Harnessing Machine Learning and Ensemble Models for Tourism Potential Zone Prediction for the Assam State of India

Harnessing Machine Learning and Ensemble Models for Tourism Potential Zone Prediction for the Assam State of India | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Harnessing Machine Learning and Ensemble Models for Tourism Potential Zone Prediction for the Assam State of India Shrinwantu Raha, Shasanka Kumar Gayen, Sayan Deb This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-4364952/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Although Assam is enriched with several popular tourist destinations but till date, its’ complete charm remains enigmatic. This research was aimed at prognosticating the Tourism Potential Zone (TPZ) for the state of Assam using five machine learning (i.e., Conditional Inference Tree, Bagged CART, Random Forest, Random Forest with Conditional Inference Tree, and Gradient Boosting models) and one ensemble model. A 5-step methodology was implemented to do this research. First, a Tourism Inventory Database was prepared using the Google earth Imagery, and a rapid field investigation carried out with the help of Global Positioning System and non-participant observation technique. Total 365 tourism points was in the inventory, 70% (224) of which was used for the training set and 30% (124) was used for the validation purpose. The tourism conditioning factors such as Relief, Aspect, Viewshed, Forest Area, Wetland, Coefficient of Variation of rainfall, Reserve Forest, Population Density, Population Growth Rate, Literacy Rate and Road-railway density were used as the independent variables in the modelling process. The TPZ was predicted with the help of above machine learning models and finally, a new TPZ Ensemble Model was proposed by combining each model. The result showed that all machine learning models performed well according to prediction accuracy and finally, the ensemble model outperformed other models by achieving the highest AUC (97.6%), Kappa (0.82) and accuracy (0.93) values. The results obtained from this research using machine learning and ensemble methods can provide proper and significant information for decision makers for the development of tourism in the region. Artificial Intelligence and Machine Learning Hospitality and Tourism Tourism potentiality Analytic Hierarchy Process ROC-AUC Conditional Inference Tree Bagged CART Random Forest TPZ Ensemble model Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Figure 8 Figure 9 Figure 10 Figure 11 Figure 12 Figure 13 Figure 14 1. Introduction In recent years, the tourism has been transpired as one of the key drivers of economic growth in both underdeveloped and wealthy nations (Manzoor et al. 2019 ). Through a number of appliances, including gains in the foreign exchange, luring in foreign investment, and raising tax receipts, tourism helps expanding the economy of both underdeveloped and developed countries (Zabihi et al. 2020 ). The economic world and its’ productivity are both facilitated by the tourism, which is one of the biggest industries in the world (Puška et al. 2021 ). Tourism is an important demographic mechanism which solves the problem of employment in both crowded and distant geographical areas. According to UNWTO’s estimates, only 25 million travellers all over the world ventured Globe-trotters in 1950. In a mega-decade, this number increased up to 1.4 billion people. In 2018, the greatest tourist growths were registered in the regions of Middle East and Asia Pacific (Scarpocchi 2020 ). The assessment of tourism potentialities is unquestionably significant and required, because tourism intrinsically and extrinsically contributes to the socioeconomic development of a place (Kachniewska 2015 ). Tourism potential is the sum of environmental, ethnographic, cultural, political and social value for establishing tourist activities in a particular place (Katelieva and Muhar 2022 ). In order to retain the location's personality, the philosophy of tourism potentiality should play a significant role (Kontogeorgopoulos 2017 ). It promotes tourists' conduct, which considerably and favorably aids in preserving the natural and chemical integrity of the ecosystem (Khadka et al. 2021 ). The tourism potentiality can evaluate a region's capacity for the sustainable and inclusive growth (Blapp and Mitas 2018 ; Zekan et al. 2022 ). As a result, it is strongly advised to evaluate the touristic potential for the integrated growth of society and culture (Banik and Mukhopadhyay 2020 ). Although the Assam state is full of tourism potential, but its’ tourism potential areas is still unexplored. Therefore, for the effective utilization of tourism resources, the demarcation and prediction of TPZ is necessary. In recent decades, the advancement of photogrammetry, GIS, and RS has led to the creation of numerous new machine learning models (ML) and algorithms by researchers worldwide (Marín-Buzón et al. 2021 ; Apostolopoulos et al. 2021). Determining TPZs involves analyzing physical and socio-economic parameters to forecast tourism trends in unexplored areas. Globally, various machine learning models are used in research fields such as identifying potential ground water zones (e.g., Vafadar et al. 2023 ; Roy et al. 2024 ), environmental and ecological planning (e.g., Mosebo Fernandes et al. 2020 ), resource management (Garg et al. 2022 ), forestry (Liu et al. 2018 ; Zhao et al. 2019 ), urban and regional planning (Chaturvedi and de Vries 2021 ; Tekouabou et al. 2022 ), and natural hazard management (Wang et al. 2021 ; Linardos et al. 2022 ). ML models were also applied for different sectors of tourism such as, tourism demand forecasting (e.g. Claveria et al. 2016 ; Bi et al. 2022 ; Karakitsiou and Mavrommati 2017 ; Cankurt and Subasi 2015 ; Law et al. 2019 ), tourist review analysis (Le et al. 2021 ; Puh and Bagić Babac 2023 ), and tourist arrival forecasting analysis (Sun et al. 2019 ), where historical time series data is available. But for the tourism potential zone identification, till date, the decision-making model particularly the Analytic Hierarchy Process (AHP) dominated the frame (e.g. Raha et al. 2021 , 2022, 2023, Sahani 2020 , Pathmanandakumar et al. 2023 ). Further, the demarcation of TPZ is always challenging because it does not necessarily rely on the time series data. Prediction of TPZ requires an active help of a fundamental GIS platform. AHP is an objective decision-making procedure, which helps to determine base information about any project over a particular region (Chowdary et al. 2013 ). Generally, Expert opinions are used to tune the model but for the similar geo-environmental factors of a same region, the models require long-term knowledge of the TPZ and its’ causal factors, which are not available in most of the cases. Several ML models play a very key role in this regard. Specifically, for the viability of the project, the advanced ML models could play a significant role by analyzing tourist behavior, accommodation and destination facilities (Singh et al. 2023 ; Nath et al. 2020 ). ML models are used to look at tourist database through pattern recognition, data driven insights, adaptability and flexibility, improved accuracy, scalability and automation (Danish et al. 2023; Chien et al. 2023 ). In this research, we have applied several machine models such as, i) Conditional Inference Tree (Kuhn and Johnson 2013) ii) Bagged CART (Hamze-Ziabari and Bakhshpoori 2018 ) iii) Random Forest Model (Gevrey et al. 2003 ) iv) Random Forest with Conditional Inference Tree (Quinlan 1992 ), and v) Gradient Boosting Models (Kuhn and Johnson 2013). Further, combining the forecasts of various boot strapping base models such as decision trees, support vector machines, bagged CART, and Gradient Boosting models, an ensemble model was prepared, which eliminate the individual model's shortcomings and benefit from their advantages. Moreover, for the spatial assessments, the ML results can be linked with GIS. This makes it possible to represent spatial data in visual form (maps), in order to identify where TPZs can be developed. This integration of the data makes results easier to understand and provides users with added spatial context during decision-making. Therefore, considering the perspectives outlined above, the research objective is to predict the TPZ using several machine learning, and ensemble models and also making an inter model comparison. The remaining manuscript was structured as follows: The second section was marked with the introduction of the study area. The third section was identified with the materials and method section. The fourth section of the manuscript identifies the ‘Result’ section of the manuscript. The fifth section highlights the ‘Discussion’ section and The last section highlights the conclusion section in the manuscript. 2. Materials and methods 2.1 Study area The Assam, the largest state in terms of people and second-largest in terms of territory, is bordered to the north by Bhutan and the Arunachal Pradesh, Nagaland, Arunachal Pradesh, and Manipur to the east, Bangladesh and Meghalaya and Tripura Mizoram and West Bengal to the south. According to Assam State Portal (assam.gov.in). Initially the Assam state had 27 districts however in 2016 about 5 districts were added however for this research and simplification total 27 districts were considered. The fact that Assam is home to three of India's six physiographic divisions—the Northern Himalayas (Eastern Hills), Northern Plains (Brahmaputra plain), and Deccan Plateau is an important geographical feature (Karbi Anglong). Assam often experiences a "Tropical Monsoon Rainforest Climate," with high humidity and significant precipitation. Warm summers and mild winters make for a temperate climate that residents may take advantage of all year round. The seasons of spring (March–April) and fall (September–October) are often pleasant with mild temperatures and rains. According to the Census of India (2011) (“Census Tables | Government of India”), Assam has a total population of 3.12 Cr. Assam's population makes up 2.58 percent of all Indians in 2011. Assam has a total population of 31,205,576 people, with 15,939,443 men and 15,266,133 women. Assam has a total size of 78,438 square kilometers. As a result, Assam has a population density that is greater than the national average. Assam is extremely important for drawing tourists because of its magnificent mountains, biodiversity, plenty of foliage, ethnic diversity (fairs and festivals), and emission zone (Huismann, 2014 ). Despite the country's popularity for tourists visiting, there are still many opportunities to explore its scenic beauty, grasslands (Bugyals), caves, bird watching sites, camping grounds, parks, and wildlife sanctuaries as well as its skiing areas, river valleys, passes, glaciers, mountain peaks, trekking trails, and river rafting locations (Choden et al. 2018 ). The Himalaya region's mountains actually give people, especially tha chance toescape the pre-monsoon heat because of the good weather and picturesque scenery (Huismann 2014 ). Till date, to the best of our knowledge, this research is the first to highlight on the tourism potentialities of the entire Assam. Previous research works highlighted the tourism potentials only in certain pockets of the state. For example, the nature-based tourism potentials of the Tinsukia District of the Upper Assam were divulged by Bordoloi and Agarwal ( 2015 ). The tourism potentials of the Karabi Anglong autonomous council districts were estimated by Ronghang and Sen (2022). But the previous research activities completely neglected the spatial assessments, model building and its’ validation. This research emphasized the tourism potentials of the Assam state using machine learning algorithms, making it an appealing reading for academicians and tourism practitioners. Therefore, this research is novel and has its’ implications on assessments of tourism potentiality. 2.2 First Step-Preparation of Tourism Inventory Database (TID) The first step of the methodology (Fig. 2 ) was started by structuring the Tourism Inventory Database (TID). TID was important to understand assess the relationship that exists between the distribution of tourism potentiality and pre-depositing factor of tourism. The tourist spots of the present study area are characterized by archaeological, natural attraction, gift shops, guest houses, hotels, memorials, monuments, museums, parks, restaurants, churches, masjid, Hindu temples, arts centre, observation towers, and stadiums. Prominent archaeological spots which were identified are Namdang Xilor Xaaku and Sri Surya Pahar. Some prominent attractions of the study region are Sualkuchi, Gateway of Assam, Sukreshwar Ghat, Kamalabari, Bhogpur Satra, Barali Range mountain etc. Prominent gift shops are Arfed, Purbashree and The Eco Hub. The Guest Houses found in the state are Niribili Guest House, Subansiri Lodge, green Red Resort, Kaziranga Guest House, Drongo Guest House etc. Some hotels and restaurants are Mayuri, Prathma, Aroma Residency, Vishal, Heritage, Appayan Restaurant, TerraMaya, Di Cafeteria, Four Season’s Restaurant etc. Children’s Memorial and Children Educational and Career Development Centre are two prominent memorials in this state. Assam State Museum, Guwahati Planetarium, State Art Gallery, and Tai Museum are very renowned in the study area. Padum Pukhuri, Trimurti Udyan, Marut Kandan, Gandhi Bagh, Kaziranga National Orchid and Biodiversity, are the few examples of parks identified in the study region. Some Monasteries in the study area are Tinsukia Buddha Vihar, Dibong Buddhist Monastery, Kumchai Kong Monastery, Namphakey Buddhist Monastery, Powai Mukh Buddhist Monastery etc. This state is also enriched with Hindu temples (i.e., Kamalabari Temple, Ghanta Temple, Shivdham Temple, Krishna Dham Temple, Nepali Temple, Maa Kamakhya Temple etc.) and churches (i.e., Dibrugarh Charch, Beno Nomaya, Northern Evangelical Lutheran Church, Baptist Church, Christ Church etc.). This section also very enriched with Pure Muslim culture and traditions. Mosques (i.e., Duliajan Mosque, Fulertal Mukam, Bongaigaon Mosque etc.) bear witness to Islamic culture. Bihu Toli is a prominent Arts Centre in the Assam. Kachujan Stadium, Maligaon Railway Stadium, Nurul Amin Stadium, Rajiv Gandhi Indoor Stadium etc. are some of the important stadiums identified in the region. This region bears a perfect tourist viewshed with the help of several observational tower such as, Charboli, Fort, Kuri Bill Camp, Latajhar etc. Figure 3 (a-m) highlights some of the tourist spots identified during the field conducted in between 19.04.2021 to 20.06.2022 in three phases. Different tourist spots were identified (latitude and longitude were collected using GPS) by the non-participant observation technique (Banik and Mukhopadhyay 2020 ). Further the official government websites were also consulted ( https://tourism.assam.gov.in/portlets ). 3.2 Second Step- Choosing tourism potentiality causative factors and multicollinearity Tourism potentiality is a multidimensional concept in which several physical and socio-ecological variables are linked into it (Raha and Gayen 2021 ). The second step of the methodology (Fig. 2 ) was started by choosing the tourism potentiality causative factors and measurement of multicollinearity. Here, 11 criteria were choosed by the recommendation of 5 expert panel (Table 1 ) and in-depth literature reviews (i.e., Raha and Gayen 2023 ; Sahani 2020 ; Trukachev 2015; Schultze et al. 2014 ; Gourabi and Rad 2013 ; Hoang et al. 2018 ). The specialists were those with a minimum of five years of expertise in the travel and tourism industry. In a separate consent form, the experts, who were consulted gave their permission for the results to be used purely for academic purposes without revealing their original identity. Criteria in this research include relief, aspect, viewshed, forest area, wetland, coefficient of variation of rainfall, the reserved forest, population density, population growth rate, literacy rate and road/railway density. The relief, aspect, forest area, wetland, coefficient of variation of rainfall and reserved forest create initial base of the tourism activities (Yuxi and Linsberg 2020). A regional population structure is highlighted through the societal indicators. Table 1 Considered thematic layers and their sources of data Criteria Sources of data Directionality of influence Descriptions Relief (RL) SRTM DEM, spatial Resolution 90 m The moderate to low relief and aspect helps to vibrate the tourism potentialities positively (Codrea et al. 2022 ). Aspect (AS) SRTM DEM, spatial Resolution 90 m Viewshed (VS) SRTM DEM, spatial Resolution 90 m If the viewshed increases, tourism potentialities also increase and vice-versa (Sahani 2020 ). Forest area (FA) India State Forest Report 2019 Forest area, wetland and rainfall variation and reserved forest positively enhance the tourism potentialities (Deribew et al. 2022 ). Wetland (WL) Assam Project On Forest and Biodiversity Conservation (Apfbc) Society https://apfbcs.nic.in/apfbcs/wetland/annexure2.pdf CV of Rainfall (CVR) Climate Research and Service Unit, Pune, (2020) Reserve Forest (RF) Assam Project On Forest and Biodiversity Conservation (Apfbc) Society https://apfbcs.nic.in/apfbcs/wetland/annexure2.pdf Population Density (PD) Statistical Hand Book of Assam, 2016 The population density and growth rate indicate the balanced population structure therefore, it motivates the positive outcomes of tourism potentiality (Fakfare et al. 2020 ) Population Growth Rate (PGR) Statistical Hand Book of Assam, 2016 Literacy Rate (LR) Statistical Hand Book of Assam, 2016 The literacy rate and the RRD both positively motivates the potential value for the tourism. Road & Railway Density (RRD) Open Street Map The sources of data were highlighted in the Table 1 . The relief, aspect and viewpoint were prepared from the Digital Elevation Model prepared by the Shuttle Radar Topographic Mission (SRTM) (Resolution is 1 ARC, pixel depth was 16 bits, published in 2014). The SRTM DEM was downloaded from the NASA Earth Explorer(“EarthExplorer,”). The information about the forest area was obtained from the India State Forest Report (2019). Information about the wetland and reserved forest were acquired from the Assam Project On Forest and Biodiversity Conservation (Apfbc) Society ( https://apfbcs.nic.in/apfbcs/wetland/annexure2.pdf ). The Coefficient of Variation of Rainfall data (2020) was obtained from the Climate Research and Service Unit, Pune, (2020). The Population Density, Population Growth Rate and Literacy Rate were obtained from the Statistical Handbook of Assam (2016). The multicollinearity is an essential prechecking before developing a particular model (Memon et al. 2019 ). A high degree of correlation between two or more dependent variables; creates disturbance in the predictability of the model. In this research, the tolerance and Variance Intrusion Factor (VIF) methods were utilized to detect it. The formula was as follows (Eq. 1 and Eq. 2 ): $$Tolerence of i-th predictor variable \left({T}_{i}\right)=1-{R}_{i}^{2}$$ 1 $$VIF of i-th predictor variable \left({T}_{i}\right)=\frac{1}{1-{R}_{i}^{2}}$$ 2 Where, \({R}_{i}^{2}\) depicts the coefficient of determination in the regression equation. For the independency, the tolerance level should be more than 0.10 and the VIF should be less than 10.0. 3.3 Third Step-Spatial modelling of TPZ In the third step (Fig. 2 ), TPZs were predicted using the single classification tree model (ctree model), Bagged CART (treebag model), Random Forest model (rf model), Random Forest with bagging ensemble algorithms utilizing conditional inference tree (cforest model) and Gradient Boosting (gbm model) models. Further, combining the ctree, treebag, rf, cforest and gbm models a new ensemble model was prepared. For this research, we have set the scenario as the 10-fold cross validation with repeats in train control. Here, the objective with cross-validated data set was to optimize and determine the size of the tree by tuning the complexity parameters. 3.3.1 Conditional Inference Tree (ctree model) The ctree model is a non-parametric class of regression trees embedding tree structured regression models into well-defined theory of conditional inference procedures. It is applicable to all kind of regression problems including multiple response scales of covariates (Hothorn et al. 2015 ; Fu 2017 ). The response Y given the status of m covariates by means of tree structured recursive partitioning. The m dimensional covariate vector \(X=({X}_{1},\dots \dots ..,{X}_{m})\) was taken from the sample space \(\aleph ={\aleph }_{1\dots \times \dots \dots \dots \dots ..}\times {\aleph }_{m.}\) Here both response and covariates could be measured in the arbitrary scales. The conditional function distribution D(Y/X) of the response Y given the covariates X depends on the function f of covariates (Eq. 3 ) (Hothorn et al. 2015 ) $$D\left(Y|X\right)=D\left(Y|{X}_{1},\dots \dots \dots \dots \dots .{X}_{m}\right)=D\left(Y\right|f\left({X}_{1,\dots \dots \dots \dots ..}{X}_{m}\right)$$ 3 Where, we restrict ourselves to portion based regression relationships i.e., r disjoint cells, \({B}_{1,\dots \dots ..}{B}_{r, }\) partitioning covariate space i.e., \(\aleph ={U}_{K=1}^{r}{B}_{K.}\) The regression relationship should be fitted on a learning sample \({L}_{n}\) i.e, learning samples with n independent observations possibly with same coordinates \({X}_{ij}\) missing (Eq. 4 ) $${L}_{n}=\{\left({Y}_{i,\dots ..}{X}_{1i,\dots \dots \dots \dots .,}{X}_{mi}\right);i=1,\dots \dots .,n\}$$ 4 A genetic algorithm for the recursive binary partitioning for a given learning sample \({L}_{n}\) can be formulated using non-negative integer valued case weights \(w=\left({w}_{1},\dots \dots \dots \dots ,{w}_{n}\right).\) Each node of the tree is attached with non-zero or zero weights as the case may be. 3.3.2 Bagged CART model (treebag model) Bagging, which stands for Bootstrap Aggregating, is a technique used to enhance the stability and accuracy of machine learning algorithms, in particular for decision trees, such as CART (treebag model) (Vrontos et al. 2021 ). This is how it worked within this research: a) Multiple bootstrap samples were created in each iteration, and each sample was essentially a random sample with replacement and is of the same size as the original dataset. B) For CART, a decision tree model was trained for each of the hundreds of bootstrap samples, which was slightly different from each other since each training sample was different in the first place. After all decision trees were trained, the predictions for the new data was made by aggregating all the predictions made by all these individual trees. For the classification tasks in this research, it was done by majority voting – the class out of all, that has the most votes. It helps to reduce the overfitting because bagging trains many models on slightly different datasets, and aggregates all their predictions together. So, it lowers the variance of the whole final model (Choi and Hur 2020 ; González et al. 2020 ). 3.3.3 Random forest model (rf model) One of the most popular supervised learning models is Random Forest (rf), which was used as a classification model in this research. Breiman ( 2001 ) first developed rf model algorithm with the help of decision trees (Zang et al. 2017; Gayen et al. 2019 ). Its’ basis is exactly the method of ensemble learning, in which a user has the opportunity to “tie” multiple classifiers together to address challenging tasks and enhance the performance of the model. One of the assets of Random Forest is essentially a decision tree. The final decision is made by result polling from different trees and selecting the most popular one. Most of the outcomes of each tree generate the end output. Random forest model provides more accuracy than the other model and also provides a clear and separate distribution plot of features in each class (Couronné, et al. 2018 ; Fox et al. 2017 ). It can handle the missing data effectively; the developed model can be saved for future use with the new data. In this research following steps were followed to build a Random Forest model: First of all, ‘K’ features from total m features were randomly selected, where k < m Node ‘d’ was calculated among ‘K’ features. Splitted the node into several daughter nodes using the best split method. Repeated the previous steps until reaching the ‘I’ number of nodes. Built a forest by repeating all steps for ‘n’ number of times to create ‘n’ number of trees. The mean squared error of each decision tree with their OOB samples (E OOB ) is used to calculate the learning error. The pros of this approach are – i) that it can deal with very voluminous data, of sufficiently large dimensions and escape the risk of overfitting (Naghibi et al. 2017 ); ii) this does not imply additional assumptions about the factors to be manipulated and the result (Youssef et al. 2016 ); iii) the analyst needs a working dataset, as there is no transformation and scalability prior to the procedure (Gayen et al. 2019 ) and iv) the model is easily applicable to any regional scale (Pourghasemi and Rahmati 2019). 3.3.4 Conditional Inference Random Forest (cforest model) The cforest (Conditional Inference Random Forest) is a combination of random forest and bagging ensemble algorithm implementation applied in this research. The main advantage of the Cforest is that it uses conditional inference trees as its base learners (Naghibi et al. 2017 ). Cforest also uses OOB data, although that means more info and better accuracy, it is also slower and can handle less data for the same memory (Thanh et al. 2022 ). The weighted average of the trees was used in this research to get the final ensemble. The main reason for the cforest’s more reliable predictions is the fact that it produces unbiased trees (Strobl et al. 2007 ; Mogensen et al. 2012 ). The cforest is always better when the model has computational resources. 3.3.5 Gradient Boosting model (gbm model) Gradient Boosting (gbm model) offer a powerful technique for tackling regression and classification problems, leveraging ensemble learning by strategically combining multiple weak algorithms into a robust solution (Islam et al. 2024 ). This method was also applied in this research. Typically employing decision trees as its basic components, this approach incrementally builds a predictive model through an iterative process, gradually boosting accuracy as each new element is incorporated into the evolving whole (Zhang et al. 2019; Sachdeva and Kumar 2021 ). The initial phase establishes a set of basic learners, which are shallow decision trees being a common choice due to their simplicity - to serve as foundations. Then, through gradient boosting, subsequent rounds fit new weak learners to the residual errors of their predecessors, reducing discrepancies between actual and predicted outcomes turn by turn. By minimizing collective mistake residuals, each additional tree tuned to those of earlier stages, the ensemble members are combined in an optimized manner that strengthens overall predictive capabilities with every included model. This model uses gradient descent optimization in order to minimize the loss function (Ridgeway 2007 ). At every iteration, it computes the gradient of the loss function with respect to the predictions of the ensemble and then it updates the predictions in the direction which minimizes the loss (Lu et al. 2020 ;. In order to fight overfitting, gbm model also uses shrinkage or learning rate. Instead of using additions of smaller number of trees, a contribution of a less than one scaled of each tree are added to the ensemble. Smaller contributions will make the optimization faster and more conservative which makes it require a higher number of iterations but will protect the model against overfitting. It can also use some kind of regularization; it can be either tree pruning which removes some of the splits producing no positive results or it can limit the maximum possible depth of the trees. 3.3.6 Ensemble model By combining the above models an ensemble model was prepared and applied for the prediction of TPZ. In this research it is regarded as ‘ TPZ Ensemble Model’ . An ensemble model is a machine learning method involving combining single models for producing a more precise prediction far more robust than any one model individually (Ganaie et al. 2022 ; Mohammed et al. 2023). The rationale for this ensemble model is consistent with the so-called “wisdom of the crowd,” meaning that the average belief of numerous models (here, ctree, treebag, rf, cforest and gbm models) is more efficient and reliable than the view of any particular model. In this research, stacking, or stacked generalization, was used to build the ‘TPZ Ensemble model’. Stacking is a very popular method, which was used in this research to prepare the ensemble model. Initially we trained base models such as, ctree, treebag, rf, cforest and gbm models. Next, instead of simply averaging or voting; another meta model or blender was introduced here. After that, the meta model was trained to learn the best way to combine the predictions of the base models to make the final prediction. During the prediction phase, the base models make their predictions on new data, and then the meta-model combines these predictions to give the final output (Polikar 2006 ; Optiz and Maclin 1999). Different base models capture different aspects or patterns within the data. By using all of them, we combined to the strengths on these models, and negate their weaknesses. It introduces diversity among the models, which helps it be less vulnerable to overfitting (Gomes et al. 2017 ). If one model overfits to some patterns within the data, others might capture different patterns or generalize better towards unseen data (Dong et al. 2020 ; Opitz and Maclin 1999 ; Polikar 2006 ). Ensemble methods are more robust to noise and outliers (Mienye and Sun 2022 ; Rokach 2010 ). For instance, if a particular model’s prediction is unreliable when the input data is outside the training data, ensembles can help improve this prediction using other model’s reliable predictions made based on that input data (Blockeel 2011 ; Rincy and Gupta 2020 ; Ganaie et al. 2022 ). In this research, all models were computed in the RStudio Version 2023.12.1 with Intel(R) Core (TM) i5-9300H CPU @ 2.40GHz 2.40 GHz processor with 8GB Ram and 64-bit operating system, 3.5 Fourth Step- accuracy assessments One of the most crucial aspects of model building is the evaluation of the output models' precision (Das 2020 ). Here, the TPZ was validated using the Kappa Coefficient, Accuracy and AUC-ROC curve. The ROC-AUC shows how specificity and sensitivity are traded off. The ROC is a two-dimensional graph in which x-axis depicts the specificity and the y axis depicts the sensitivity. The euqations 5, 6 and 7 illustrated the attributes of x and y axis, where, the TN represents true negative, FP represents false positive, TP represents true positive, and FN represents false negative (Eq. 5 , Eq. 6 and Eq. 7 ) (Roodposhti et al. 2017 ): $$x=specificity=\left[\frac{TN}{(TN+FP)}\right]$$ 5 $$y=sensitivity=\left[\frac{TP}{(TP+FN)}\right]$$ 6 $$Accuracy=\left[\frac{TP+TN}{(TP+TN+FP+FN)}\right]$$ 7 The Kappa coefficient was defined as (Eq. 8 ) $$Kappa=\left[\frac{{P}_{0}-{P}_{est}}{1-{P}_{est}}\right]$$ 8 Where, \({P}_{0}\) is defined as the observed agreement; and \({P}_{est}\) is the expected agreement. The performance of the model is quantitatively depicted by the area under the ROC curve (AUC) (Tang et al. 2020 ). A standard scale of AUC is i) ≥ 0.9 denotes excellent ii) 0.8 to 0.9 denotes accepted iii) 0.7 to upto 0.8 denotes good or satisfactory, iv) 0.5 to 0.7 is considerable and v) less than 0.5 is rejected (Trabelsi et al. 2023 ; Mitra et al. 2022 ). It is recommended that the machine learning models should be judged based on the validation or test dataset (Vabalas et al. 2019 ). Because the validation sets are commonly used for hyperparameter tuning where different hyperparameters configurations of the model are tested to find the best-performing one. This ensures that the model’s performance is optimized for the specific dataset while still keeping its ability to generalize. 4. Results and discussion 4.1 Analysis of multicollinearity For all criteria, the tolerance level fluctuated from .322 (for population density) to .977 (for aspect) and the VIF varied from 1.217 (for reserved forest) to 2.839 (for forest area) (Table 2 ). The highest VIF assures the lowest tolerance level. Here, all VIF values are less than 10 and all tolerance level values are less than 1.0. Therefore, this research shows that multicollinearity is absent in the eleven pre-processing factors being investigated. Table 2 Tolerance level and Variance Intrusion Factor for different indicators Criteria Tolerance VIF RL .491 2.036 AS .977 1.023 CVR .495 2.022 FA .352 2.839 VS .595 1.681 RF .822 1.217 WL .756 1.323 PGR .496 2.017 LR .458 2.182 PD .322 3.102 RRD .814 1.229 4.2 Spatial interrelationship between tourism location and causative factors Distribution of tourism locations and causative factors were illustrated in this research using Analytic Hierarchy Process (AHP) model. The AHP is an unbiased MCDM method for choosing the best option from a large pool of alternatives (Munier and Hontoria 2021 ; Senapati and Das 2021 ). The AHP method was invented by Saaty, ( 1980 , 1987 ), and it attracted a wide number of researchers for its’ adaptability and usefulness. First, the pair-wise comparative matrix, which represents the relative priorities of each criterion, has an identical number of rows and columns was prepared. The 5-member expert panel chose the significance of several criteria in this case. The panel was created by including those experts and researchers, who had at least five years of experience in the field of travel and tourism. The experts who were consulted agreed, in a separate consent form, that the scores might be used only for academic purposes, without disclosing their identities. The preference of each criterion was estimated using a relative dominance scale of 1 to 9 (Saaty 1980 ). Here, 1, 3, 5, 7 and 9 were marked as equal importance, moderate importance, strong importance, very strong importance and extreme or substantial importance. 2,4,6 and 8 represent intermediate values. The necessary condition of a considerable AHP matrix is that the consistency ratio or C.R. value should be less than 0.1, The relief of Assam varied from 1 meter to 1971meter (Fig. 4 a). The Western and the Eastern sections of the study area were with lower elevation (i.e.,1 meter to 300 meters). The southern sections are comparatively high (i.e., 300 meters to 1971 meters). The tourism potentiality decreases with the increasing relief value. As a result, priority rises as the class value of the relief decrease, and vice versa. 1 to 300-meter relief class had higher areal coverage (87.055% area). There also an inverse trend between the very high relief and tourist locations. Therefore, Tourism potentiality pixels are substantially overlapped within the moderate to low relief classes. The aspect fluctuated from − 1 to 359.458 (Fig. 4 b). Lower aspect is getting more attention in case of tourism and its potentiality. In this study, the AS was divided into four categories and importance increased as class value decreased. The first category (-1 to 89.211) was directed towards flat, northern, north-eastern and eastern direction. The Bramhaputra river passes in the middle portion of the Assam and along the river lowest aspect value was marked (1 to 89.211). The second category (89.212 to 179.424) was marked towards the eastern, south-eastern and southern directions. The third category (179.424 to 269.636) was directed towards the southern, south-west and western directions. The fourth category (269.637 to 359.848) was marked towards the western, northwestern and northern directions. As the distance increases from the river, the aspect value increases. Tourism potentiality pixels are largely overlapped in moderate to low aspect values (i.e., 194 pixels). The forest area fluctuated from 4.52–86.07% within the study area. Larger forest area was marked in the districts of Karbi Anglong, West Karbi Anglong, Dima Hasao, Cachar, Hailakandi, and Karimganj. 30–50% forest area was found in the Kakrajhar, Chirang, Golpara, Kamrup, Karimganj and Tinsukia districts. Remaining districts were marked with comparatively low forest area (i.e., 4.52–30% area). The picturesque attractiveness is positively accelerated by the FA, which draws many adventure travelers (Karali et al. 2021 ). The FA was splitted into four groups for this study, and as the category value is higher, weightages likewise increase and vice versa (Fig. 4 c). Moderate to high forest cover attracts more tourism potentiality pixels (i.e., 162 pixels). The number of reserved forests in the Assam fluctuated from 0 to 29 (Fig. 7 d). Comparatively higher number of reserved forests (i.e., 16 to 29 numbers) were found in the West Karabi Arlong, Naogaon, Karbi Arlong, Tinsukia and Kamrup Metropolitan districts. Remaining districts were marked with lower number (0 to 15 numbers) of reserved forests. The number of protected forestry has a favorable impact on the tourism potential. Here, in this research, the RF was classified into four classes and the class value increases, weightages also increase and vice-versa (Fig. 4 d). Moderate and high classes of RF incur higher tourism potentiality pixels in the study area (i.e., 234 pixels). The number of wetlands in the study area varied from 0 to 1790 (Fig. 5 a). The higher number of wetlands (i.e., 251 to 1790 numbers) were identified in the districts of Kakrajhar, Karbi Arlong, Naogaon, West Karbi Arlong, Sontipur, Hailakandi, Karimganj and Tinsukia districts. The remaining portions were noticed with lower number of wetlands. The WL favorably vibrates the tourism potentiality. Here, the WL was broken down into four classes, and priority increases with the increasing class value. Tourism potentiality pixels also substantially merged under very high, high and moderate class values of WL (i.e., 228 pixels). The coefficient of rainfall variation (CVR) fluctuated from 86.401–108.007% (Fig. 5 b). Most parts of the Assam experienced a large variation of rainfall. Comparatively higher amount of rainfall variation (92.889–108.006%) was noticed in the Dhubri, Kokrajhar, Golpara, Bongajagaon, Barpeta, Chirang, Baksa, Nalbari, Kamrup, Karabi Arlong, Darang, Kamrup Metropolitan, Morigaon, Sontipur, Lakhimpur, Dhemji, Naogaon, Dima Hasaao, Cachar, Hailakandi and Kamrup districts. The remaining portions were noticed with lower amount of rainfall variation. The CVR positively enhances the potential for tourism (Giorgi and Lionello 2008 ). Four categories were used to classify the CV in this instance, and as the value of each category increases, so does priority, and vice versa. Tourism location pixels also substantially merged under very high, high and moderate class values of CVR (i.e., 228 pixels). The viewshed of the study area was marked by identifying 17 major hills (Fig. 5 c). Here, the viewshed was classified into 4 classes. The Western sections of the study area were more prominent. The VS positively enhances the potential for tourism (Giorgi and Lionello 2008 ; Ferguson et al. 2022). The VS was split into four categories in this instance, and priority increases as the class value increases. Tourism potentiality pixels are overlapped majorly over very high, high and moderate classes of VS (i.e. 273 pixels). The PGR of the research area varied from 5.210–24.440% (Fig. 6 a). Comparatively high PGR value (i.e., 20 to 24%,) was found in the Dhubri, Bongaigaon, Golpara, Barpata, Darrang, Kamrup Metropolitan, Morigaon, Cacchar, Hailakandi and Karimganj districts. Remaining districts have 5.210–20% growth rate. Upper Assam has comparatively low growth rate. The PGR in this instance was separated into four categories, and priority increases as the quantity of each class increases and vice versa. High, Very High and Moderate classes of PGR are identified with a higher number of pixels counts of tourism potentiality (i.e., 311 pixels).The PD of the study area varied from 44 to 1313 persons per square km. Comparatively high PD value (i.e., 701 to 1313 persons/sq.km,) was found in the Dhubri, Barpata, Nalbari, Kamrup Metropolitan and Naogaon districts. Remaining districts have 44 to 700 persons /square km (Fig. 6 b). High population density negatively affects the tourism potentiality. Therefore, the PD in this instance was divided into four separate classes, and importance increases as the class value decreases and vice versa. The substantial tourism potentiality pixels overlapped in the low and moderate PD classes (i.e., 315 pixels). The LR varied from 65.37–88.71% within the study area of Assam. Comparatively low literacy rates were marked in the Dhubri, Chirang, Sontipur, Baksa, Golpara, Jorhat and Tinsukia districts. Higher literacy rates were identified in the Darrang, Kamrup Metropolitan, Morigaon, Golaghat, Jorhat, Sivsagar, Karbi Anglong, Hailakandi, and Karimganj districts (Fig. 6 c). As tourism is a tertiary activity, therefore, the LR would positively vibrate the tourism activities. Here, the LR was classified into four groups, and as the value of each class rises, so does priority, and vice versa. High, very high and moderate classes of LR combinedly achieve higher number of pixels for tourism potentiality (i.e., 310 pixels). Road and railway networks connect the tourism destinations quickly and smoothly (Bast et al. 2016 ). As a result, the RRD was divided into four categories. As the class value rises, so does priority, and vice versa. Moderate, high and very high classes of RRD combinedly overlapped with higher tourism potentiality pixel value (i.e., 249 pixels). The detailed areal coverage of different thematic layer, their classes and weightages were marked in Table 3 . Table 3 Relationship between the variables with the help of the AHP method Indicators Class Category Priority % (Weightage) Principal eigen value & Consistency Ratio (C.R.) Number of Pixels in the domain Number of tourism potentiality pixels in the domain % Area RL 1.00 - 80.00 (Low) 80.01 - 300.00 (Moderate) 300.01 - 800.00 (High) 800.01 - 1971.00 (Very High) 1.00 - 80.00 1 3 4 5 (55.00%) 8 Principal eigen value = 4.057, Consistency Ratio CR = 2.1% 3696933 140 39.234 80.01 - 300.00 0.33 1 2 2 (21.40) % 6 44485020 178 47.281 300.01 - 800.00 0.25 0.5 1 2 (14.20%) 4 9533666 37 10.249 800.01 - 1971.00 0.2 0.5 0.5 1 (9.40%) 2 2508725 10 2.697 AS -1 - 89.211 (Low) 89.212 - 179.424 (Moderate) 179.425 - 269.636 (High) 269.637 - 359.848 (Very High) -1 - 89.211 1 2 5 7 (54.50%) 8 Principal eigen value = 4.072, Consistency Ratio CR = 2.6% 23798802 80 25.583 89.212 - 179.424 0.5 1 2 3 (24.70%) 6 24023779 114 25.825 179.425 - 269.636 0.2 0.5 1 3 (14.10%) 5 22381165 95 24.059 269.637 - 359.848 0.14 0.33 0.33 1 (6.70%) 3 22820598 73 24.532 FA 50.001 - 86.070 (Very high) 30.001- 50.00 (High) 15.001 - 30.00 (Moderate) 4.520 - 15.000 (Low) 50.001 - 86.070 1 2 5 7 (53.20%) 8 Principal eigen value = 4.118, Consistency Ratio CR = 4.3% 7858 101 25.856 30.001 - 50.00 0.5 1 3 3 (27.30%) 6 5793 62 19.062 15.001 - 30.000 0.2 0.33 1 3 (12.80%) 5 12149 134 39.976 4.520 - 15.000 0.14 0.33 0.33 1 (6.70%) 3 4591 66 15.106 WL 501 – 1790 (very high) 251 – 500 (High) 101 – 250 (Moderate) 0 – 100 (Low) 501 - 1790 1 3 4 7 (56.30%) 8 Principal eigen value = 4.063, Consistency Ratio CR = 2.3% 2616 30 8.608 251 - 500 0.33 1 2 3 (22.30%) 6 11800 142 38.827 101 - 250 0.25 0.5 1 3 (14.80%) 4 4737 56 15.587 0 - 100 0.14 0.33 0.33 1 (6.70%)3 11238 134 36.978 CVR 86.401 - 87.563 (Low) 87.564 - 92.888 (Moderate) 92.889 - 101.239 (High) 101.240 - 108.006 (Very high) 86.401 - 87.563 1 2 4 7 (52.50%) 7 Principal eigen value = 4.050, Consistency Ratio CR = 1.8% 4293 87 14.126 87.564 - 92.888 0.5 1 2 3 (25.40%) 6 6896 124 22.691 92.889 - 101.239 0.25 0.5 1 3 (15.20%) 5 8816 78 29.009 101.240 - 108.006 0.14 0.33 0.33 1 (6.90%) 3 10386 73 34.175 VS 7.1 – 12 (Very High) 4.1 – 7 (High) 1.1 - 4 (Moderate) 0 – 1 (Low) 7.1 - 12 1 3 4 8 (57.20%) 8 Principal eigen value = 4.052, Consistency Ratio CR = 1.9% 5051 60 16.669 4.1 - 7 0.33 1 2 3 (22.00%) 7 2932 37 9.676 1.1 - 4 0.25 0.5 1 3 (14.50%) 5 8614 176 28.428 0 - 1 0.12 0.33 0.33 1 (6.30%) 3 13704 92 45.226 RF 23-29 (very high) 16-22 (High) 8-15 (Moderate) 0-7 (Low) 23-29 1 3 4 8 (56.40%) 8 Principal eigen value = 4.048, Consistency Ratio CR = 1.7% 2788 31 9.174 16-22 0.33 1 2 5 (24.20%) 6 2258 26 7.430 8-15 0.25 0.5 1 3 (13.90%) 4 14813 177 48.741 0-7 0.12 0.2 0.33 1 (5.40%) 3 10532 128 34.655 PD 44 – 350 (Low) 351 – 700 (Moderate) 701 – 1000 (High) 1001 – 1313 (Very High) 44 - 350 1 2 4 5 (50.70%) 8 Principal eigen value = 4.021, Consistency Ratio CR = 0.8% 10642 108 35.017 351 - 700 0.5 1 2 3 (26.40%) 6 15875 207 51.940 701 - 1000 0.25 0.5 1 2 (14.30%) 4 3651 43 12.013 1001 - 1313 0.2 0.33 0.5 1 (8.60%) 3 313 04 1.030 PGR 20.001 - 24.440 (Very high) 15.001 - 20.000 (High) 10.001 - 15.000 (Moderate) 5.210 - 10.000 (Low) 20.001 - 24.440 1 2 4 5 (51.20%) 8 Principal eigen value = 4.047, Consistency Ratio CR = 1.7% 8213 98 27.024 15.001 - 20.000 0.5 1 2 2 (24.40%) 6 11272 138 37.090 10.001 - 15.000 0.25 0.5 1 2 (14.60%) 4 6737 75 22.168 5.210 - 10.000 0.2 0.5 0.5 1 (9.80%) 2 4169 51 13.718 LR 79.31-88.71 (Very high) 72.64-79.30 (High) 65.38-72.63 (Moderate) 58.34-65.37 (Low) 79.31-88.71 1 2 3 5 (48.80%)8 Principal eigen value = 4.041, Consistency Ratio CR = 1.5% 313 47 1.030 72.64-79.30 0.5 1 2 2 (25.20%) 6 4704 116 15.478 65.38-72.63 0.33 0.5 1 2 (16.10%) 4 8539 147 28.097 58.34-65.37 0.2 0.5 0.5 1 (10.00%) 2 16835 50 55.395 RRD 300.01 - 482.69 (Very High) 70.01 - 300.00 (High) 30.01 - 70.00 (Moderate) 0 - 30.00 (Low) 300.01 - 482.69 1 2 3 9 (52.70%) 8 Principal eigen value = 4.021, Consistency Ratio CR = 0.8% 156 01 0.513 70.01 - 300.00 0.5 1 1 3 (21.50%) 6 1337 16 4.399 30.01 - 70.00 0.33 1 1 3 (19.30%) 5 9701 232 31.921 0 - 30.00 0.11 0.33 0.33 1 (6.40%) 3 19197 113 63.167 4.3 Analysis of different pre-requisites of machine learning models Before initialization of machine learning models, we develop 10-fold repeated cross validation framework. Our objective with the cross validation is to optimize the size of the tree. After plotting the 1-P value threshold versus accuracy (repeated cross validation) we observe that how accuracy is maximized into a relatively less complex tree (Fig. 7 ). The final value used for the model was mincriterion 0.01; which was used to best tune the model. In the final model total 16 terminal nodes were created in the regression trees (Fig. 8 ). Rules were illustrated for this ctree model in Appendix. 1. Within 16 terminal nodes, the model predicted criterion 1 (Very High to High TPZ) 9 times. Final Bagged CART (treebag model) model was created by using 25 bootstrap replications. For the random forest model (rf) the final accuracy was used to select the optimal model using the largest value. The final value for the rf model is mtry = 11 (best tuned). For this final model rf model, number of trees are 500 and the number of variables tried to each split is 6. The out of the bag error (OOB) estimate for the model is 17.25%. For the cforest model, the final accuracy was used to select the optimal model using the largest value mtry = 11 (best tuned) with the 500 number of trees in the final model. The final tuned gradient boosted model was created with Bernoulli Loss function with 50 iterations. It has a 50 number of trees, interaction depth 2, shrinkage 0.1 and minobsinnode 10. 4.4 Variable importance analysis Tourism Potentiality is a multifaceted conception, which are dependent on several criteria (Raha and Gayen 2021 ). Therefore, it is really essential to determine tourism potentiality effective factors and their contribution. After train each model, Variable Importance Plot (Fig. 9 ) for each model was assessed. For the ctree model (Fig. 9 a), the FA bears the highest importance. The FA was followed by LR, WL, RF, RL, CVR, RRD, AS, VS and PGR (least importance). For the treebag model (Fig. 9 b), RRD was marked with the highest importance and the PD was identified with the least importance. RRD was followed by LR, FA, RL, AS, CVR, VS, RF, PGR, WL and PD. In the rf model (as shown in Fig. 9 c), the feature RRD was identified as having the greatest importance, with FA, LR, RL, AS, VS, PD, RF, CVR, WL, and PGR following in descending order of importance. For the cforest model (Fig. 9 d), FA was marked with the highest importance followed by RRD, LR, PD, RF, RL, PGR, CVR, VS, WL and AS. The FA was also found with the highest importance for the gbm model (Fig. 9 e). Here WL was found with the least importance. RRD and LR were also found with substantial importance for the gbm model. For the TPZ Ensemble Model (Fig. 9 f), rf model was found the highest importance followed by gbm, treebag, ctree and cforest model. 4.5 Tourism Potentiality assessment models The tourism potentiality assessment models by applying five machine learning (i.e., ctree, treebag, rf, cforest, and gbm) and one ensemble model were presented in this research (Fig. 10 a, 10 b, 11 a, 11 b, 12 a, 12 b). The all models show very high, high and moderate to low tourism potentiality followed by natural break strategy (Gayen et al. 2019 ). Higher value indicates higher tourism potentiality. Using Ensemble model approximately, 35.42% area was marked as the high to very high tourism potentiality. On he other hand, approximately 64.58% area was identified as the moderate to low tourism potentiality. Using the Bagged CART model (treebag), approximately 53.61% area was identified as the high to very high tourism potentiality and 46.39% area was demarcated as moderate to low tourism potentiality. Using the Conditional Inference Tree (ctree) model, approximately 50.62% area was demarcated as high to very high tourism potentiality and remaining 49.38% area was marked as the moderate to low tourism potentiality. Using Conditional Inference Random Forest (cforest) model, near about 40.52% and 59.48% area were identified as the high to very high and moderate to low tourism potentiality respectively. Using the rf model, approximately 28.24% area was identified as very high to high tourism potentiality. Remaining 71.76% area was identified with moderate to low tourism potentialities. Approximately 36.5% area was demarcated as High to Very High tourism potentialities by the gradient boosting (gbm model) model. On the contrary, 63.5% area was marked with moderate to low tourism potentiality (Table 4 ). Table 4 Pixel count with percentage area by different models Pixel Count % Area Models Moderate to Low High Very High Total Moderate to Low High Very High Total Ensemble 19339 8353 2253 29945 64.58 27.89 7.53 100 Bagged CART 13891 13152 2902 29945 46.39 43.92 9.69 100 Conditional Inference Random Forest 17811 11608 526 29945 59.48 38.76 1.76 100 Conditional Inference Tree 14787 11716 3442 29945 49.38 39.13 11.49 100 Random Forest 21489 7350 1106 29945 71.76 24.55 3.69 100 Gradient Boosting 19015 10488 442 29945 63.5 35.02 1.48 100 4.4.1 Very High (VH) and High (H) TPZ The detailed characteristics of this zone was briefly discussed as follows: VH and H tourism potentialities dominated in the Middle portions, south-eastern and southern sections of the Assam. Over 50% area of Golpara, Hailakandi, Jorhat, Kamrup, Kamrup Metropolitan, Karbi Anglong, Karimganj, Lakhimpur, Naogaon, Sivasagar, Tinsukia and West Karbi Anglong districts were marked with Very high potentials for tourism (Fig. 10 a, 10 b, 11 a, 11 b, 12 a, 12 b). These sections were marked with the moderate to high relief structure (80.01 meter to 1971 meter) (Fig. 4 a), which creates a wider viewshed of the location and this attracts tourists conveniently and easily (Huff and Tingley 2015 ). Higher number of reserved forests (8 to 29 numbers, Fig. 4 d) and larger forest area (30.001–86.07% area, Fig. 4 c) in these sections created amazing ambience for the tourism activities. It is the region, which is dominated by the wetlands (i.e., 251 to 1790 Fig. 5 a) and rainfall variation (i.e., 92.889–108.006%, Fig. 5 b). A wetland is an area where the ground is continuously or periodically flooded by water, whether it be salty, pure, or a combination of both. The seasonal variation of rainfall accelerates the vegetation pattern and forest cover of a particular region. The seasonal variation creates the rthymic diversity of forest cover (Ghazoul 2016 ). This section has a higher road and rail density (greater than 70, Fig. 6 d), which accelerates the connectivity of the region. Tourists can reach the tourist destination more relatedly and easily by accessible roads and rails (Holloway and Humphreys 2022 ). The population density (351 to 1000 persons/ sq.km., Fig. 6 b), population growth rate (10–24%, Fig. 9 a) and the literacy rate (72.64–88.71%, Fig. 9 c) are higher in these sections of the study area. The literacy rate creates awareness about a balanced population structure; therefore, it creates favorable environment for the tourism activities (Getz and Page 2019 ). Irrespective of the above issues, these portions are affected by comparatively low pollution levels, as it is dominated by higher number of reserved forests and forest area (Singh et al. 2020 ). As a result, these portions are marked with higher potentials for the tourism. 4.4.2 Moderate (M) to Low (L) TPZ The detailed characteristics of this zone was briefly discussed as follows: The upper Assam and the upper north-eastern portions were marked with the low tourism potentials. Over 50% area of Baksa, Barpeta, Bongaigaon, Chirang, Darrang, Dhemaji, Dhubri, Dibrugarh, Golaghat, Kokrajhar, Morigaon, Nalbari, Sontipur, and Udayguri were identified under the moderate to low tourism potentials (Fig. 10 a, 10 b, 11 a, 11 b, 12 a, 12 b). Comparatively flat terrain (1 to 300meter, Fig. 4 a, 4 b, 4 c) was marked in these sections which does not create any picturesque beauty and do create a low viewshed of a particular region. Near about 5–30% forest area (Fig. 4 d) and comparatively low number of reserved forests (0 to 15 numbers). These factors lower the ambience, vibrations and motivations of the tourist activities. These sections are scarced of wetlands (i.e., 0 to 250 numbers Fig. 5 a) and these sections are marked with a very low amount of rainfall variation (i.e., 86.401–92.888%, Fig. 5 b). Moderate to high population density (i.e., 351 to 1000 persons/square km., Fig. 6 b) and growth rate (i.e., 10.001–20.000%) are also marked in these portions of the study area (Fig. 6 a). Literacy rates (i.e., 58.34–72.63%, Fig. 6 c) are also comparatively poor. These portions have low to moderate RRD values (0 to 70) (Fig. 6 d). Apart from the above specified issues, these portions are affected by comparatively high pollution levels, as it is crowded with different types of industries. These sections are over congested. Therefore, these sections are not attracted by tourists hence having a comparatively low tourism potential. 4.3 Validation The Ensemble model was appeared with the highest Kappa (0.81), accuracy (0.93 value) and AUC-ROC (96.9% area) values. Based on AUC-ROC measurement, the Ensemble model was followed by cforest model (89.7% AUC), rf model (79.9% AUC), gbm model (78.8% AUC values), ctree model (74.9% AUC values) and treebag model (74.01% AUC-ROC). According to the quality criteria of AUC, the performance of ensemble model was appeared as ‘excellent’. The performance of cforest model was ‘accepted’. The performance of remaining models (i.e., ctree, treebag, gbm, and rf models) are ‘good’ or ‘satisfactory’. Based on the accuracy and Kappa measurement, also Ensemble model outperformed other models. The Ensemble model was followed by rf model, gbm model, cforest model, ctree model and treebag model. For both cases, ctree and treebag are the worst performer. On the contrary, rf, cforest and gbm models performed well. ROC-AUC and accuracy plots were outlined in Fig. 13 and Fig. 14 and Table 5 . Table 5 Accuracy of different models (predictive accuracy) Model names Accuracy Kappa Rank (Based on Accuracy and Kappa) Model Names AUC_ROC Values 95% Confidence Interval (For AUC) Rank (Based on AUC) Upper Bound Lower Bound Ensemble 0.92 0.81 1 Ensemble 96.9% 0.850 0.962 1 Conditional Inference Tree (cforest) 0.81 0.62 4 Conditional Inference Tree (cforest) 89.7% 0.723 0.878 2 Random Forest (rf) 0.85 0.68 2 Random Forest (rf) 79.9% 0.713 0.870 3 Stochastic Gradient Boosting (gbm) 0.84 0.67 3 Stochastic Gradient Boosting (gbm) 78.8% 0.693 0.854 4 Conditional Inference Tree (ctree) 0.80 0.60 5 Conditional Inference Tree (ctree) 74.9% 0.806 0.896 5 Bagged CART (treebag) 0.79 0.57 6 Bagged CART (treebag) 74.01% 0.693 0.895 6 5. Discussion The Ensemble method is the most accurate approach used in this research, according to the results of accuracy assessments. Combining the forecasts of various boot strapping base models such as decision trees, random forests, gradient boosting, ensemble methods eliminated the individual model's shortcomings and benefit from their advantages. Here the ensemble method helps to improve predictive performance by drawing on the combined wisdom of multiple models. Therefore, the formulation of ‘TPZ ensemble model’ enlarges a new direction in tourism research. In a move to be increasingly precise and effective, meticulous work was done on a new ensemble model of wide variety machine learning methodologies. This product, the TPZ ensemble model, has outstripped not only individual models but it is pure predictive ability in both the unparalleled accuracy metrics. Rigorous scrutiny showed that the TPZ ensemble model surpassed its constituent parts significantly, with an exceptional Area Under the Curve (AUC) value for example 96.9%. Also, its strong performance is evident in a Kappa coefficient of 0.81 - substantial agreement beyond pure chance. The most amazing thing is that the ensemble model, with its generalization capability still intact, actually gets 92% right. Further, the Ensemble methods helps in the decision-making process and helps choosing the best model from a variety and bunch of several ML models. Further, the rf and cforest models handled big datasets without variable deletion may have contributed to their strong performance in this study. According to Catani et al. ( 2013 ), the rf model can handle nonlinearities between dominant factors and therefore provide good performance in the current context. With respectable performances, this model has also shown to be helpful in other research domains, including mapping groundwater potential, predicting wildfires, modeling sediment yield (Masselink et al., 2017 ), and mapping landslide susceptibility (LSM) (Taalab et al. 2018 ; Zhang et al. 2017 ). The Gradient boosting Models also performed well in prediction accuracy may be attributed to the fact that it combines the predictions of several base estimators, typically decision trees. From the current study, the accuracy of the combined models was much higher compared to that of any individual model and also because decision trees are not sensitive to outliers. The high performance is scaled further by their ability to support large datasets and therefore can be used in applications with huge datasets. However, decision trees (like the Conditional Inference Tree and the Bagged CART models) have the ability to isolate outliers in different leaves, which keeps them from having a substantial impact on the performance of the model as a whole. Also, large data cannot be efficiently handled by ctree and Bagged CART model. They are also sensitive to outliers. Although in this research, their performance is quite satisfactory but in comparison to other models, the prediction accuracy, Kappa and AUC-ROC is quite low. Conclusion In this research, an ensemble model and a variety of machine learning algorithms (i.e., Conditional Inference Tree, Bagged CART, Random Forest, Random Forest with Conditional Inference Tree, Gradient Boosting and one ensemble model) were used to predict the Tourism Potential Zones (TPZ) for the state of Assam. Here, first of all, we created a comprehensive Tourism Inventory Database from field research and Google Earth imagery, which produced 365 tourism points. Our study produced encouraging results by utilizing a wide range of tourism conditioning factors as independent variables, such as Relief, Aspect, Viewshed, Forest Area, Wetland, and socio-economic criteria like Population Density, Literacy Rate, and Road-Railway Density. The results clearly showed the good quality of the maps generated, with the best agreement between the ML models and tourism inventory data points. All of the spatially assessed TPZ maps show that Very High to high tourism potentialities dominated in the south-eastern and southern sections of the Assam. High to very high tourism potentials in this research are associated with moderate to low relief, higher viewshed, forest area, wetland, rainfall variation, higher literacy rate and better communication network (here Road and Railway density). Over 50% area of Golpara, Hailakandi, Jorhat, Kamrup, Kamrup Metropolitan, Karbi Anglong, Karimganj, Lakhimpur, Naogaon, Sivasagar, Tinsukia and West Karbi Anglong districts were marked with very high to high tourism potentials and Over 50% area of Baksa, Barpeta, Bongaigaon, Chirang, Darrang, Dhemaji, Dhubri, Dibrugarh, Golaghat, Kokrajhar, Morigaon, Nalbari, Sontipur, and Udayguri were identified under the lower tourism potentials. We found that all models performed admirably in terms of prediction accuracy after conducting a thorough analysis using Kappa, Accuracy and AUC-ROC method. TPZ ensemble model was proposed by combining other base models, and it is interesting to note that the ensemble model emerged as the frontrunner, showcasing superior metrics including the highest AUC (97.6%), Kappa (0.82), and accuracy (0.93) values. This research underscores the potential of machine learning and ensemble methods in predicting TPZs, furnishing valuable insights for decision-makers tasked with spearheading the development of tourism in Assam. By offering robust and nuanced predictions, our findings contribute to informed decision-making processes aimed at harnessing the rich tourism potential of the region, thereby fostering sustainable growth and prosperity. Declarations Acknowledgement Authors expresses their gratitude to all faculty members of Department of Geography, Bhairab Ganguly College, who wholeheartedly support this research. Conflict of Interest There is no conflict of interest regarding publication of this article. Funding Source This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. Compliance with Ethical Standards The ethical approval is not applicable for this research. Data availability statement Data are available from authors upon a reasonable request. References Apostolopoulos, D, Nikolakopoulos, K (2021) A review and meta-analysis of remote sensing data, GIS methods, materials and indices used for monitoring the coastline evolution over the last twenty years. Eur J Remote Sens 54:240-265. https://doi.org/10.1080/22797254.2021.1904293 Banik, S, Mukhopadhyay, M (2020) Model-based strategic planning for the development of community-based tourism: a case study of Ayodhya Hills in West Bengal, India. GeoJournal:1-17. https://doi.org/10.1007/s10708-020-10314-0 Bast, H, Delling, D, Goldberg, A, Müller-Hannemann, M, Pajor, T, Sanders, P, … Werneck R F (2016) Route Planning in Transportation Networks. In L. Kliemann & P. Sanders (Eds.), Algorithm Engineering: Selected Results and Surveys (pp. 19–80). Cham: Springer International Publishing. https://doi.org/10.1007/978-3-319-49487-6_2 Bi JW, Han TY, Li H (2022) International tourism demand forecasting with machine learning models: The power of the number of lagged inputs. Tourism Econ 28:621-645. https://doi.org/10.1177/1354816620976954 Blapp M, Mitas, O (2018) Creative tourism in Balinese rural communities. Current Issues in Tourism 21:1285–1311. https://doi.org/10.1080/13683500.2017.1358701 Blockeel H (2011) Hypothesis Space Encyclopedia of Machine Learning, 511–513. Bordoloi AK, Agarwal BK (2015) Tourism Potentiality in Tinsukia district of Upper Assam: An Analysis. International Journal of Management and Development Studies 4:375–383. https://doi.org/10.53983/ijmds.v4i4.164 Breiman L (2001) Random forests. Machine learning 45:5-32. Cankurt S, Subasi A (2015) Developing tourism demand forecasting models using machine learning techniques with trend, seasonal, and cyclic components. Balkan Journal of Electrical and Computer Engineering 3:42-49. https://dergipark.org.tr/en/download/article-file/39709 Catani F, Lagomarsino D, Segoni S, Tofani V (2013) Landslide susceptibility estimation by random forests technique: sensitivity and scaling issues. Nat Hazard Earth Syst Sciences 13:2815-2831. Chaturvedi V, de Vries WT. (2021) Machine learning algorithms for urban land use planning: A review. Urban Science 5:68. https://doi.org/10.3390/urbansci5030068 Chien C F, Ku C C, Lu Y Y (2023) Ensemble learning for demand forecast of After-Market spare parts to empower data-driven value chain and an empirical study. Computers & Industrial Engineering 185:109670. https://doi.org/10.1016/j.cie.2023.109670 Choden K, Wangchuk D, Smart! C (2018) Bhutan—Culture Smart! The Essential Guide to Customs & Culture. Kuperard. Choi S, Hur J (2020) An ensemble learner-based bagging model using past output data for photovoltaic forecasting. Energies 13:1438. Chowdary V M, Chakraborthy, D, Jeyaram, A, Murthy Y K, Sharma J R, Dadhwal VK (2013) Multi-criteria decision-making approach for watershed prioritization using analytic hierarchy process technique and GIS. Water Rsour Manage 27: 3555-3571. https://doi.org/10.1007/s11269-013-0364-6 Claveria O, Monte E, Torra S (2016) Combination forecasts of tourism demand with machine learning models. Appl Econ Letter 23:428-431. https://doi.org/10.1080/13504851.2015.1078441 Codrea PM, Bilașco Ș, Roșca S, Irimuș IA, Iuliu V, Rusu R, ... Sestras P (2022) The Integrated Assessment of Degraded Tourist Geomorphosites to Develop Sustainable Tourism: A Case Study of Grădina Zmeilor Geomorphosite, North-West Region, Romania. Appl Sci-Basel 12: 9816. https://doi.org/10.3390/app12199816 Couronné R, Probst P, Boulesteix AL (2018) Random forest versus logistic regression: a large-scale benchmark experiment. BMC Bioinformatics 19:1-14. https://doi.org/10.1186/s12859-018-2264-5 Danish M S S (2023) AI and expert insights for sustainable energy future. Energies 16: 3309. https://doi.org/10.3390/en16083309 Das S (2020) Flood susceptibility mapping of the Western Ghat coastal belt using multi-source geospatial data and analytical hierarchy process (AHP). Remote Sensing Applications: Society and Environment 20:100379. https://doi.org/10.1016/j.rsase.2020.100379 Deribew KT, Mihretu Y, Abreha G, Gemeda DO (2022) Spatial analysis of potential ecological sites in the northeastern parts of Ethiopia using multi-criteria decision-making models. Asia-Pacific Journal of Regional Science 6: 961–991. https://doi.org/10.1007/s41685-022-00248-5 Dong X, Yu Z, Cao W, Shi Y, Ma Q (2020) A survey on ensemble learning. Front Comput Sci-Chi 14:241-258. Fakfare P, Talawanich S, Wattanacharoensil W (2020) A scale development and validation on domestic tourists’ motivation: The case of second-tier tourism destinations. Asia Pacific Journal of Tourism Research, 25:489–504. https://doi.org/10.1080/10941665.2020.1745855 Fox EW, Hill RA., Leibowitz SG., Olsen AR., Thornbrugh DJ., Weber MH. (2017) Assessing the accuracy and stability of variable selection methods for random forest modeling in ecology. Environ Monit Assess 189:1-20. https://doi.org/10.1007/s10661-017-6025-0 Fu W (2017) Nonparametric Methods in Statistical Learning: Unbiasedness in Regression Trees, Survival Trees for Nonstandard Data and Estimating the Number of Clusters (Doctoral dissertation, New York University, Graduate School of Business Administration). Ganaie MA, Hu M, Malik AK., Tanveer M, Suganthan PN (2022) Ensemble deep learning: A review. Eng Appl of Artif Intel 115:105151. https://doi.org/10.1016/j.engappai.2022.105151 Gayen A, Pourghasemi HR, Saha S, Keesstra S, Bai S (2019) Gully erosion susceptibility assessment and management of hazard-prone areas in India using different machine learning algorithms. Sci Total Environ, 668:124-138. Getz D, Page SJ (2019) Theory, Research and Policy for Planned Events (4th ed.). London: Routledge. https://doi.org/10.4324/9780429023002 Gevrey M, Dimopoulos I, Lek S (2003) Review and comparison of methods to study the contribution of variables in artificial neural network models. Ecol Model 160:249-264. Ghazoul J (2016) Dipterocarp Biology, Ecology, and Conservation. Oxford University Press. Garg S, Sinha S, Kar AK, Mani M (2022) A review of machine learning applications in human resource management. International Journal of Productivity and Performance Management 71:1590-1610. https://doi.org/10.1108/IJPPM-08-2020-0427 Giorgi, F, Lionello P (2008) Climate change projections for the Mediterranean region. Global Planet Change 63:90–104. https://doi.org/10.1016/j.gloplacha.2007.09.005 Gomes HM, Barddal JP, Enembreck F, Bifet A (2017) A survey on ensemble learning for data stream classification. ACM Comput Surv (CSUR) 50:1-36. González S, García S, Del Ser J, Rokach L, Herrera F (2020) A practical tutorial on bagging and boosting based ensembles for machine learning: Algorithms, software tools, performance study, practical perspectives and opportunities. Inform Fusion 64:205-237. https://doi.org/10.1016/j.inffus.2020.07.007 Gourabi B R, Rad TG (2013) The analysis of ecotourism potential in Boujagh wetland with AHP method. Life Science Journal 10: 251-258. Hamze-Ziabari S M, Bakhshpoori T (2018) Improving the prediction of ground motion parameters based on an efficient bagging ensemble model of M5′ and CART algorithms. Appl Soft Comput, 68:147-161. https://doi.org/10.1016/j.asoc.2018.03.052 Hoang HT, Truong QH, Nguyen AT, Hens L (2018) Multicriteria evaluation of tourism potential in the central highlands of Vietnam: Combining geographic information system (GIS), analytic hierarchy process (AHP) and principal component analysis (PCA). Sustainability 10:3097. https://doi.org/10.3390/su10093097 Holloway JC, Humphreys C (2022) The Business of Tourism. SAGE. Hothorn T, Hornik K, Zeileis, A (2015) ctree: Conditional inference trees. The comprehensive R archive network 8: 1-33. https://apfbcs.nic.in/apfbcs/wetland/annexure2.pdf https://tourism.assam.gov.in/portlets Huff C, Tingley D (2015) Who are these people? Evaluating the demographic characteristics and political preferences of MTurk survey respondents. Research & Politics 2: 2053168015604648. https://doi.org/10.1177/2053168015604648. Huismann W (2014) PandaLeaks: The Dark Side of the WWF. Nordbook. Islam MT, Ayon EH, Ghosh BP, MD SC, Shahid R, Rahman S, ... Nguyen TN (2024) Revolutionizing Retail: A Hybrid Machine Learning Approach for Precision Demand Forecasting and Strategic Decision-Making in Global Commerce. Journal of Computer Science and Technology Studies 6:33-39. https://doi.org/10.32996/jcsts.2024.6.1.4 Kachniewska MA (2015) Tourism development as a determinant of quality of life in rural areas. Worldwide Hospitality and Tourism Themes 7:500–515. https://doi.org/10.1108/WHATT-06-2015-0028 Karakitsiou A, Mavrommati A (2017) Machine learning methods in tourism demand forecasting: Some evidence from Greece. MIBES transactions 11:92-105. https://www.researchgate.net/publication/323427035_Machine_learning_methods_in_tourism_demand_forecasting_some_evidence_from_Greece Karali A, Das S, Roy H (2021) Forty years of the rural tourism research: Reviewing the trend, pattern and future agenda. Tourism Recreation Research 49:173-200. https://doi.org/10.1080/02508281.2021.1961065. Katelieva M, Muhar A (2022) Heritage tourism products based on traditional nature-related knowledge: Assessment of cultural, social, and environmental factors in cases from rural Austria. Journal of Heritage Tourism 17:631–647. https://doi.org/10.1080/1743873X.2022.2098040 Khadka D, Chaudhary A, Karki R, Sharma B, Bhatta S (2021) Ecotourism in Ghoda Ghodi Wetland Sukhad, Kailali, Nepal. Journal of Tourism and Hospitality Education 11:22–42. https://doi.org/10.3126/jthe.v11i0.38237 Kontogeorgopoulos N (2017) Finding oneself while discovering others: An existential perspective on volunteer tourism in Thailand. Ann Tourism Res 65:1–12. https://doi.org/10.1016/j.annals.2017.04.006 Kuhn, Johnson (2013) Applied Predictive Modeling, Springer, New York (chapter 4). Law R, Li G, Fong DKC, Han X (2019) Tourism demand forecasting: A deep learning approach. Ann Tourism Res 75:410-423. https://doi.org/10.1016/j.annals.2019.01.014 Le TH, Arcodia C, Novais MA, Kralj A (2021) Proposing a systematic approach for integrating traditional research methods into machine learning in text analytics in tourism and hospitality. Curr Issues Tour 24:1640-1655. https://doi.org/10.1080/13683500.2020.1829568 Linardos V, Drakaki M, Tzionas P, Karnavas YL (2022) Machine learning in disaster management: recent developments in methods and applications. Machine Learning and Knowledge Extraction, 4:446-473. https://doi.org/10.3390/make4020020 Liu Z, Peng C, Work T, Candau JN., DesRochers A. Kneeshaw D (2018) Application of machine-learning methods in forest ecology: recent progress and future challenges. Environ Rev 26: 339-350. https://doi.org/10.1139/er-2018-0034 Lu H, Karimireddy SP, Ponomareva N, Mirrokni V (2020) Accelerating gradient boosting machines. In International Conference on Artificial Intelligence and Statistics :516-526. PMLR. Manzoor F, Wei L, Asif M, Haq MZ ul, Rehman H ur. (2019) The Contribution of Sustainable Tourism to Economic Growth and Employment in Pakistan. Int J Environ Res Pub He 16:3785. https://doi.org/10.3390/ijerph16193785. Marín-Buzón C, Pérez-Romero A, López-Castro J L, Ben Jerbania I, Manzano-Agugliaro F (2021) Photogrammetry as a new scientific tool in archaeology: Worldwide research trends. Sustainability 13:5319. https://doi.org/10.3390/su13095319 Masselink R, Temme A JAM, Giménez R, Casalí J, Keesstra SD (2017) Assessing hillslope-channel connectivity in an agricultural catchment using rare-earth oxide tracers and random forests models. Cuadernos De Investigación Geográfica 43:17–39. https://doi.org/10.18172/cig.3169 Memon MA, Cheah JH, Ramayah T, Ting H, Chuah F, Cham TH. (2019) Moderation analysis: issues and guidelines. Journal of Applied Structural Equation Modeling 3: 1-11. Mogensen UB, Ishwaran H, Gerds TA (2012) Evaluating random forests for survival analysis using prediction error curves. J Stat Softw 50:1. https://doi.org/10.18637%2Fjss.v050.i11 Mohammed A, Kora R (2023) A comprehensive review on ensemble deep learning: Opportunities and challenges. Journal of King Saud University-Computer and Information Sciences 35:757-774. Mosebo Fernandes AC, Quintero Gonzalez R, Lenihan-Clarke MA, Leslie Trotter EF, Jokar Arsanjani J (2020) Machine learning for conservation planning in a changing climate. Sustainability 12: 7657. https://doi.org/10.3390/su12187657 Mienye ID, Sun Y (2022) A survey of ensemble learning: Concepts, algorithms, applications, and prospects. IEEE Access 10:99129-99149. Mitra R, Saha P, Das J (2022) Assessment of the performance of GIS-based analytical hierarchical process (AHP) approach for flood modelling in Uttar Dinajpur district of West Bengal, India. Geomatics, Natural Hazards and Risk 13:2183-2226. https://doi.org/10.1080/19475705.2022.2112094 Munier, N., Hontoria, E. (2021) Uses and Limitations of the AHP Method. Cham: Springer International Publishing. Naghibi SA, Ahmadi K, Daneshi A (2017) Application of support vector machine, random forest, and genetic algorithm optimized random forest models in groundwater potential mapping. Water Resour Manag 31:2761-2775. https://doi.org/10.1007/s11269-017-1660-3 Nath B, Wang Z, Ge Y, Islam K, P Singh, R., Niu, Z (2020) Land use and land cover change modeling and future potential landscape risk assessment using Markov-CA model and analytical hierarchy process. ISPRS Int J Geo-Inf 9:134. https://doi.org/10.3390/ijgi9020134 Opitz D, Maclin R (1999) Popular ensemble methods: An empirical study. J Artif Intel Res. 11: 169–198. Pathmanandakumar V, Goh HC, Chenoli SN (2023) Identifying potential zones for ecotourism development in Batticaloa district of Sri Lanka using the gis-based Ahp spatial analysis. GeoJournal of Tourism and Geosites 46:252-261. https://doi.org/10.30892/gtg.46128-1022 Polikar R (2006) Ensemble based systems in decision making. IEEE Circ Syst Mag 6:21-45. Pourghasemi HR, Rahmati O (2018) Prediction of the landslide susceptibility: Which algorithm, which precision? Catena 162:177-192. Puh K, Bagić Babac M (2023) Predicting sentiment and rating of tourist reviews using machine learning. Journal of hospitality and tourism insights 6:1188-1204. https://doi.org/10.1108/JHTI-02-2022-0078 Puška A, Pamucar D, Stojanović I, Cavallaro F, Kaklauskas A, Mardani A (2021) Examination of the Sustainable Rural Tourism Potential of the Brčko District of Bosnia and Herzegovina Using a Fuzzy Approach Based on Group Decision Making. Sustainability 13: 583. https://doi.org/10.3390/su13020583 Quinlan JR (1992) Learning with continuous classes. In 5th Australian joint conference on artificial intelligence 92:343-348). Raha S, Gayen SK (2023) Application of Analytic Hierarchy Process and weighted sum techniques for green tourism potential mapping in the Gangetic West Bengal, India. GeoJournal 88(Suppl 1): 197-240. https://doi.org/10.1007/s10708-022-10619-2 Raha S, Gayen SK (2021) Tourism Potentiality Zone Mapping by Using the AHP Technique: A Study on Bankura District, West Bengal, India. Journal of Geographical Studies 6: 58-85. https://doi.org/10.21523/gcj5.22060201 Raha S, Mondal M, Gayen SK (2021) Ecotourism potential zone mapping by using analytic hierarchy process (AHP) and weighted linear algorithm: A study on West Bengal, India. Journal of Geographical Studies 5: 44-64. https://doi.org/10.21523/gcj5.21050201 Ridgeway G (2007) Generalized Boosted Models: A guide to the gbm package. Update, 1: 2007. Rincy TN, Gupta R (2020) Ensemble learning techniques and its efficiency in machine learning: A survey. In 2nd international conference on data, engineering and applications (IDEA) IEEE pp:1-6. Rokach L (2010) Ensemble-based classifiers. Artificial Intelligence Review 33: 1–39 Ronghang S, Sen Dr S (2022) Tourism potentials in the Karbi Anglong autonomous council districts (KAAC) of Assam. International Journal of Geography, Geology and Environment 4:112–117. https://doi.org/10.22271/27067483.2022.v4.i2b.123. Roodposhti MS, Safarrad T, Shahabi H (2017) Drought sensitivity mapping using two one-class support vector machine algorithms. Atmos Res 193:73-82. https://doi.org/10.1016/j.atmosres.2017.04.017 Roy SK, Hasan MM, Mondal I, Akhter J, Roy SK, Talukder S, ... Karuppannan S (2024) Empowered machine learning algorithm to identify sustainable groundwater potential zone map in Jashore District, Bangladesh. Groundwater for Sustainable Development 25: 101168. https://doi.org/10.1016/j.gsd.2024.101168 Sachdeva S, Kumar B (2021) Comparison of gradient boosted decision trees and random forest for groundwater potential mapping in Dholpur (Rajasthan), India. Stoc Env Res Risk A 35:287-306. https://doi.org/10.1007/s00477-020-01891-0 Sahani N (2020) Application of analytical hierarchy process and GIS for ecotourism potentiality mapping in Kullu District, Himachal Pradesh, India. Environment, Development and Sustainability 22: 6187-6211. https://doi.org/10.1007/s10668-019-00470-w Saaty TL (1980) The analytic hierarchy process: Planning, priority setting, resource allocation. McGraw Hill. Saaty, RW (1987) The analytic hierarchy process—What it is and how it is used. Math Modelling 9:161–176. https://doi.org/10.1016/0270-0255(87)90473-8. Scarpocchi C (2020) Where are people going? In Place Branding . Routledge. Schultze, J, Gärtner S, Bauhus J, Meyer P, Reif A (2014) Criteria to evaluate the conservation value of strictly protected forest reserves in Central Europe. Biodivers Conserv 23:3519-3542. Senapati U, Das TK (2021) Assessment of basin-scale groundwater potentiality mapping in drought-prone upper Dwarakeshwar River basin, West Bengal, India, using GIS-based AHP techniques. Arab J Geosci 14:960. https://doi.org/10.1007/s12517-021-07316-8 Singh KR, Goswami AP, Kalamdhad AS, Kumar B (2020) Assessment of surface water quality of Pagladia, Beki and Kolong rivers (Assam, India) using multivariate statistical techniques. International Journal of River Basin Management 18: 511–520. https://doi.org/10.1080/15715124.2019.1566236 Singh S, Rao MJ, Baranval NK, Kumar, KV, Kumar YV (2023) Geoenvironment factors guided coastal urban growth prospect (UGP) delineation using heuristic and machine learning models. Ocean Coast Manage. 236:106496. https://doi.org/10.1016/j.ocecoaman.2023.106496 Strobl C, Boulesteix AL, Zeileis A, Hothorn T (2007) Bias in random forest variable importance measures: Illustrations, sources and a solution. BMC Bioinformatics 8:1-21. Sun S, Wei Y, Tsui KL, Wang S (2019) Forecasting tourist arrivals with machine learning and internet search index. Tourism Manage 70:1-10. https://doi.org/10.1016/j.tourman.2018.07.010 Taalab K, Cheng T, Zhang Y (2018) Mapping landslide susceptibility and types using Random Forest. Big Earth Data 2:159-178. Tang Z, Zhao W, Xie X, Zhong Z, Shi F, Liu J, Shen D (2020) Severity assessment of coronavirus disease 2019 (COVID-19) using quantitative features from chest CT images. arXiv preprint arXiv:2003.11988. https://doi.org/10.48550/arXiv.2003.11988 Tekouabou SCK, Diop EB, Azmi R, Jaligot R, Chenal J (2022) Reviewing the application of machine learning methods to model urban form indicators in planning decision support systems: Potential, issues and challenges. Journal of King Saud University-Computer and Information Sciences 34: 5943-5967. https://doi.org/10.1016/j.jksuci.2021.08.007 Thanh NN, Chotpantarat S, Trung NH., Ngu NH (2022) Mapping groundwater potential zones in Kanchanaburi Province, Thailand by integrating of analytic hierarchy process, frequency ratio, and random forest. Ecol Indic 145:109591. https://doi.org/10.1016/j.ecolind.2022.109591 Trabelsi, F, Bel Hadj Ali S, Lee S (2023) Comparison of Novel Hybrid and Benchmark Machine Learning Algorithms to Predict Groundwater Potentiality: Case of a Drought-Prone Region of Medjerda Basin, Northern Tunisia. Remote Sens Basel 15:152. https://doi.org/10.3390/rs15010152 Trukhachev A (2015) Methodology for evaluating the rural tourism potentials: A tool to ensure sustainable development of rural settlements. Sustainability 7: 3052-3070. Youssef AM, Pourghasemi HR, Pourtaghi ZS, Al-Katheeri MM (2016) Erratum to: Landslide susceptibility mapping using random forest, boosted regression tree, classification and regression tree, and general linear models and comparison of their performance at Wadi Tayyah Basin, Asir Region, Saudi Arabia. Landslides, 13, 1315-1318. Vabalas A, Gowen E, Poliakoff E, Casson AJ (2019) Machine learning algorithm validation with a limited sample size. PloS one 14: e0224365. https://doi.org/10.1371/journal.pone.0224365 Vafadar S, Rahimzadegan M, Asadi R (2023) Evaluating the performance of machine learning methods and Geographic Information System (GIS) in identifying groundwater potential zones in Tehran-Karaj plain, Iran. J Hydrol 624:129952. https://doi.org/10.1016/j.jhydrol.2023.129952 Vrontos SD, Galakis J, Vrontos ID. (2021) Modeling and predicting US recessions using machine learning techniques. Int J Forecasting 37: 647-671. https://doi.org/10.1016/j.ijforecast.2020.08.005 Wang C, Yu Q, Law KH, McKenna F, Stella XY., Taciroglu E, ... Cetiner B (2021) Machine learning-based regional scale intelligent modeling of building information for natural hazard risk management. Automat Constr, 122:103474. https://doi.org/10.1016/j.autcon.2020.103474 Yuxi, Z, Linsheng, Z. (2020) Identifying conflicts tendency between nature-based tourism development and ecological protection in China. Ecol Indic, 109:105791. Zabihi H, Alizadeh M, Wolf ID, Karami M, Ahmad A, Salamian H (2020) A GIS-based fuzzy-analytic hierarchy process (F-AHP) for ecotourism suitability decision making: A case study of Babol in Iran. Tourism Management Perspectives 36:100726. https://doi.org/10.1016/j.tmp.2020.100726 Zekan B, Weismayer C, Gunter U, Schuh B, Sedlacek S (2022) Regional sustainability and tourism carrying capacities. J Clean Prod 339:130624. https://doi.org/10.1016/j.jclepro.2022.130624 Zhang K, Wu X, Niu R, Yang K, Zhao L (2017) The assessment of landslide susceptibility mapping using random forest and decision tree methods in the Three Gorges Reservoir area, China. Environ Earth Sci 76:1-20. Zhao Q, Yu S, Zhao F, Tian L, Zhao Z (2019) Comparison of machine learning algorithms for forest parameter estimations and application for forest quality assessments. Forest Ecol Manag 434:224-234. https://doi.org/10.1016/j.foreco.2018.12.019. Additional Declarations The authors declare no competing interests. Supplementary Files Appendix.docx Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-4364952","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":298460791,"identity":"42c2c557-b7e2-4dac-9aa4-e6edb1034459","order_by":0,"name":"Shrinwantu Raha","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA/0lEQVRIiWNgGAWjYDACCQaGAzA2M5ArB2IceECMFh6IFgtjsJYEAloYkLRUJDaAePi0yM/ufXi4sM0uz14i9+HnghqJ9Plhhx8CbbGT023ArsXgznGDwzPbkot5JNKNpWcck8jdeDvNAKgl2djsAA4tEmkMh3m3MSf2ABnSPGxALbMTQFoOJG7DoUV+BlhLPUgL82+efxLphrPTP+DVwnADrOUwSAubNG+bRIK8dA5+WwzAWv4dT+w584zNmrdPwnCDdE7BgQQD3H4BOoz5M8+Z6sT29jTm2zzf6uTlZ6dv/vChwk4OlxYs9oJVGhCrHGxvAymqR8EoGAWjYCQAADNKXVtE8HonAAAAAElFTkSuQmCC","orcid":"","institution":"Bhairab Ganguly College","correspondingAuthor":true,"submittingAuthor":false,"prefix":"","firstName":"Shrinwantu","middleName":"","lastName":"Raha","suffix":""},{"id":298460792,"identity":"05ff9ce7-f29f-47d2-a2f5-2e555d6f9b9d","order_by":1,"name":"Shasanka Kumar Gayen","email":"","orcid":"","institution":"Cooch Behar Panchanan Barma University","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Shasanka","middleName":"Kumar","lastName":"Gayen","suffix":""},{"id":298460793,"identity":"dd93ae1d-d9c3-4c02-a1ca-ec0d48dc4d11","order_by":2,"name":"Sayan Deb","email":"","orcid":"","institution":"Bhairab Ganguly College","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Sayan","middleName":"","lastName":"Deb","suffix":""}],"badges":[],"createdAt":"2024-05-03 15:24:38","currentVersionCode":1,"declarations":{"humanSubjects":false,"vertebrateSubjects":true,"conflictsOfInterestStatement":false,"humanSubjectEthicalGuidelines":false,"humanSubjectConsent":false,"humanSubjectClinicalTrial":false,"humanSubjectCaseReport":false,"vertebrateSubjectEthicalGuidelines":true,"coiExplicitlySet":false},"doi":"10.21203/rs.3.rs-4364952/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-4364952/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":56025077,"identity":"c140f029-fc71-49e3-a307-92331c4737a5","added_by":"auto","created_at":"2024-05-07 16:55:48","extension":"jpeg","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":593915,"visible":true,"origin":"","legend":"\u003cp\u003eLocation map\u003c/p\u003e","description":"","filename":"image1.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-4364952/v1/6285214090bc616d7cd79cab.jpeg"},{"id":56025987,"identity":"bcf7cd2d-8207-498b-9484-617194ae704d","added_by":"auto","created_at":"2024-05-07 17:03:49","extension":"jpeg","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":247597,"visible":true,"origin":"","legend":"\u003cp\u003eMethodological Framework\u003c/p\u003e","description":"","filename":"image2.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-4364952/v1/7a8b4e130e65b4eeaa86f5a3.jpeg"},{"id":56025084,"identity":"fe0d212a-cf86-4a05-9be8-c3c9d7201743","added_by":"auto","created_at":"2024-05-07 16:55:49","extension":"jpeg","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":860716,"visible":true,"origin":"","legend":"\u003cp\u003e(a-m) Field photographs of different tourist spots in Assam\u003c/p\u003e","description":"","filename":"image3.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-4364952/v1/34391fd763b9c58955fd3b16.jpeg"},{"id":56025082,"identity":"9caa637c-3cf3-4486-9fe0-ba9ae4ff91c8","added_by":"auto","created_at":"2024-05-07 16:55:49","extension":"jpeg","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":431400,"visible":true,"origin":"","legend":"\u003cp\u003eCriteria of the study area; a) Relief b) Aspect c) % of forest area d) Number of reserved forests\u003c/p\u003e","description":"","filename":"image4.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-4364952/v1/7ca187e636e32a0a4174726d.jpeg"},{"id":56025081,"identity":"7e41f34f-0b8a-4867-9ec5-55472cc59c4b","added_by":"auto","created_at":"2024-05-07 16:55:49","extension":"jpeg","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":327881,"visible":true,"origin":"","legend":"\u003cp\u003eCriteria of the study area; a) Number of wetland b) Coefficient of Variation of rainfall c) Viewshed\u003c/p\u003e","description":"","filename":"image5.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-4364952/v1/63f1b287ddaecc91e1859d30.jpeg"},{"id":56025080,"identity":"ed3530ed-7e1e-43be-96dc-30e4a5ba403c","added_by":"auto","created_at":"2024-05-07 16:55:49","extension":"jpeg","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":343838,"visible":true,"origin":"","legend":"\u003cp\u003eCriteria of the study area; a) Growth Rate (%), b) Population Density c) Literacy Rate (%) d) Road and Railway density\u003c/p\u003e","description":"","filename":"image6.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-4364952/v1/3b7890b55c6295c9931af204.jpeg"},{"id":56025083,"identity":"13eb0bbf-1b5a-4810-8003-43774188c025","added_by":"auto","created_at":"2024-05-07 16:55:49","extension":"jpeg","order_by":7,"title":"Figure 7","display":"","copyAsset":false,"role":"figure","size":102065,"visible":true,"origin":"","legend":"\u003cp\u003eThreshold vs. Accuracy Plot\u003c/p\u003e","description":"","filename":"image7.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-4364952/v1/2fd6ef2363376d6b5f16457c.jpeg"},{"id":56025087,"identity":"59e2fbb4-3122-4362-8ca7-9bd0874da8e3","added_by":"auto","created_at":"2024-05-07 16:55:49","extension":"jpeg","order_by":8,"title":"Figure 8","display":"","copyAsset":false,"role":"figure","size":244099,"visible":true,"origin":"","legend":"\u003cp\u003eRegression tree for Conditional Inference Tree (ctree) model (HW- High to Moderate class; LW – Low Class\u003c/p\u003e","description":"","filename":"image8.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-4364952/v1/5a2b1035766fe85e5d73edf8.jpeg"},{"id":56025990,"identity":"687eeded-049c-4d69-bd05-8d6cf3eb8f9f","added_by":"auto","created_at":"2024-05-07 17:03:50","extension":"jpeg","order_by":9,"title":"Figure 9","display":"","copyAsset":false,"role":"figure","size":229152,"visible":true,"origin":"","legend":"\u003cp\u003eVariable Importance Plot of different models; a) ctree model b) treebag model c) rf model d) cforest model e) gbm model f) ensemble model\u003c/p\u003e","description":"","filename":"image9.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-4364952/v1/7e1c0a56e0d670b5ebeab7c8.jpeg"},{"id":56025088,"identity":"27143356-4e5f-4310-a194-0b39b41dfc6f","added_by":"auto","created_at":"2024-05-07 16:55:49","extension":"png","order_by":10,"title":"Figure 10","display":"","copyAsset":false,"role":"figure","size":728009,"visible":true,"origin":"","legend":"\u003cp\u003ea) Tourism Potential Zone (TPZ) predicted by a) Conditional Inference Tree Model (ctree model); b) Bagged CART model (treebag model)\u003c/p\u003e","description":"","filename":"image10.png","url":"https://assets-eu.researchsquare.com/files/rs-4364952/v1/eb456ad32639eafcd73ce3a9.png"},{"id":56025989,"identity":"fabf0d14-f394-4f60-8e90-316746f2b88d","added_by":"auto","created_at":"2024-05-07 17:03:49","extension":"png","order_by":11,"title":"Figure 11","display":"","copyAsset":false,"role":"figure","size":706241,"visible":true,"origin":"","legend":"\u003cp\u003eTourism Potential Zone (TPZ) predicted by a) rf model; b) cforest model\u003c/p\u003e","description":"","filename":"image11.png","url":"https://assets-eu.researchsquare.com/files/rs-4364952/v1/dbbca7644c1a618a11cc0194.png"},{"id":56025090,"identity":"6fed07a2-b0f0-4427-8ed2-716c89c5a26e","added_by":"auto","created_at":"2024-05-07 16:55:49","extension":"png","order_by":12,"title":"Figure 12","display":"","copyAsset":false,"role":"figure","size":531177,"visible":true,"origin":"","legend":"\u003cp\u003eTourism Potential Zone (TPZ) predicted by a) gbm model; b) Ensemble model\u003c/p\u003e","description":"","filename":"image12.png","url":"https://assets-eu.researchsquare.com/files/rs-4364952/v1/04863272e7f86b29a8f3114b.png"},{"id":56025086,"identity":"6e35a930-b3ea-4307-8d96-d3f37fd9b6ba","added_by":"auto","created_at":"2024-05-07 16:55:49","extension":"jpeg","order_by":13,"title":"Figure 13","display":"","copyAsset":false,"role":"figure","size":155033,"visible":true,"origin":"","legend":"\u003cp\u003eAUC_ROC by different machine learning models (Detailed results are stored in Table 6)\u003c/p\u003e","description":"","filename":"image13.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-4364952/v1/bf58d3fde7be091f00080f7a.jpeg"},{"id":56025078,"identity":"739a2fc2-d07c-4280-9217-03f0918a1b03","added_by":"auto","created_at":"2024-05-07 16:55:49","extension":"jpeg","order_by":14,"title":"Figure 14","display":"","copyAsset":false,"role":"figure","size":109003,"visible":true,"origin":"","legend":"\u003cp\u003eAccuracy and Kappa plots for different models\u003c/p\u003e","description":"","filename":"image14.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-4364952/v1/460f4bd08a0318594fdd5936.jpeg"},{"id":56026765,"identity":"791041c9-ca87-49d4-aa3d-0a2823f687eb","added_by":"auto","created_at":"2024-05-07 17:11:52","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":4295109,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-4364952/v1/b877bac5-2848-44a5-806a-dbfe3a024bdc.pdf"},{"id":56025076,"identity":"73868709-655f-4ef9-a595-06a8806b46bb","added_by":"auto","created_at":"2024-05-07 16:55:48","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":13900,"visible":true,"origin":"","legend":"","description":"","filename":"Appendix.docx","url":"https://assets-eu.researchsquare.com/files/rs-4364952/v1/e440fe145d16389b2eb0c722.docx"}],"financialInterests":"The authors declare no competing interests.","formattedTitle":"\u003cp\u003e\u003cstrong\u003eHarnessing Machine Learning and Ensemble Models for Tourism Potential Zone Prediction for the Assam State of India\u003c/strong\u003e\u003c/p\u003e","fulltext":[{"header":"1. Introduction","content":"\u003cp\u003eIn recent years, the tourism has been transpired as one of the key drivers of economic growth in both underdeveloped and wealthy nations (Manzoor et al. \u003cspan citationid=\"CR57\" class=\"CitationRef\"\u003e2019\u003c/span\u003e). Through a number of appliances, including gains in the foreign exchange, luring in foreign investment, and raising tax receipts, tourism helps expanding the economy of both underdeveloped and developed countries (Zabihi et al. \u003cspan citationid=\"CR108\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). The economic world and its\u0026rsquo; productivity are both facilitated by the tourism, which is one of the biggest industries in the world (Puška et al. \u003cspan citationid=\"CR74\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). Tourism is an important demographic mechanism which solves the problem of employment in both crowded and distant geographical areas. According to UNWTO\u0026rsquo;s estimates, only 25\u0026nbsp;million travellers all over the world ventured Globe-trotters in 1950. In a mega-decade, this number increased up to 1.4\u0026nbsp;billion people. In 2018, the greatest tourist growths were registered in the regions of Middle East and Asia Pacific (Scarpocchi \u003cspan citationid=\"CR89\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). The assessment of tourism potentialities is unquestionably significant and required, because tourism intrinsically and extrinsically contributes to the socioeconomic development of a place (Kachniewska \u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e2015\u003c/span\u003e). Tourism potential is the sum of environmental, ethnographic, cultural, political and social value for establishing tourist activities in a particular place (Katelieva and Muhar \u003cspan citationid=\"CR48\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). In order to retain the location's personality, the philosophy of tourism potentiality should play a significant role (Kontogeorgopoulos \u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e2017\u003c/span\u003e). It promotes tourists' conduct, which considerably and favorably aids in preserving the natural and chemical integrity of the ecosystem (Khadka et al. \u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). The tourism potentiality can evaluate a region's capacity for the sustainable and inclusive growth (Blapp and Mitas \u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e2018\u003c/span\u003e; Zekan et al. \u003cspan citationid=\"CR109\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). As a result, it is strongly advised to evaluate the touristic potential for the integrated growth of society and culture (Banik and Mukhopadhyay \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). Although the Assam state is full of tourism potential, but its\u0026rsquo; tourism potential areas is still unexplored. Therefore, for the effective utilization of tourism resources, the demarcation and prediction of TPZ is necessary.\u003c/p\u003e \u003cp\u003eIn recent decades, the advancement of photogrammetry, GIS, and RS has led to the creation of numerous new machine learning models (ML) and algorithms by researchers worldwide (Mar\u0026iacute;n-Buz\u0026oacute;n et al. \u003cspan citationid=\"CR58\" class=\"CitationRef\"\u003e2021\u003c/span\u003e; Apostolopoulos et al. 2021). Determining TPZs involves analyzing physical and socio-economic parameters to forecast tourism trends in unexplored areas. Globally, various machine learning models are used in research fields such as identifying potential ground water zones (e.g., Vafadar et al. \u003cspan citationid=\"CR104\" class=\"CitationRef\"\u003e2023\u003c/span\u003e; Roy et al. \u003cspan citationid=\"CR84\" class=\"CitationRef\"\u003e2024\u003c/span\u003e), environmental and ecological planning (e.g., Mosebo Fernandes et al. \u003cspan citationid=\"CR63\" class=\"CitationRef\"\u003e2020\u003c/span\u003e), resource management (Garg et al. \u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e2022\u003c/span\u003e), forestry (Liu et al. \u003cspan citationid=\"CR55\" class=\"CitationRef\"\u003e2018\u003c/span\u003e; Zhao et al. \u003cspan citationid=\"CR111\" class=\"CitationRef\"\u003e2019\u003c/span\u003e), urban and regional planning (Chaturvedi and de Vries \u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e2021\u003c/span\u003e; Tekouabou et al. \u003cspan citationid=\"CR98\" class=\"CitationRef\"\u003e2022\u003c/span\u003e), and natural hazard management (Wang et al. \u003cspan citationid=\"CR106\" class=\"CitationRef\"\u003e2021\u003c/span\u003e; Linardos et al. \u003cspan citationid=\"CR54\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). ML models were also applied for different sectors of tourism such as, tourism demand forecasting (e.g. Claveria et al. \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e2016\u003c/span\u003e; Bi et al. \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e2022\u003c/span\u003e; Karakitsiou and Mavrommati \u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e2017\u003c/span\u003e; Cankurt and Subasi \u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e2015\u003c/span\u003e; Law et al. \u003cspan citationid=\"CR52\" class=\"CitationRef\"\u003e2019\u003c/span\u003e), tourist review analysis (Le et al. \u003cspan citationid=\"CR53\" class=\"CitationRef\"\u003e2021\u003c/span\u003e; Puh and Bagić Babac \u003cspan citationid=\"CR73\" class=\"CitationRef\"\u003e2023\u003c/span\u003e), and tourist arrival forecasting analysis (Sun et al. \u003cspan citationid=\"CR95\" class=\"CitationRef\"\u003e2019\u003c/span\u003e), where historical time series data is available. But for the tourism potential zone identification, till date, the decision-making model particularly the Analytic Hierarchy Process (AHP) dominated the frame (e.g. Raha et al. \u003cspan citationid=\"CR78\" class=\"CitationRef\"\u003e2021\u003c/span\u003e, 2022, 2023, Sahani \u003cspan citationid=\"CR86\" class=\"CitationRef\"\u003e2020\u003c/span\u003e, Pathmanandakumar et al. \u003cspan citationid=\"CR70\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). Further, the demarcation of TPZ is always challenging because it does not necessarily rely on the time series data. Prediction of TPZ requires an active help of a fundamental GIS platform. AHP is an objective decision-making procedure, which helps to determine base information about any project over a particular region (Chowdary et al. \u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e2013\u003c/span\u003e). Generally, Expert opinions are used to tune the model but for the similar geo-environmental factors of a same region, the models require long-term knowledge of the TPZ and its\u0026rsquo; causal factors, which are not available in most of the cases. Several ML models play a very key role in this regard. Specifically, for the viability of the project, the advanced ML models could play a significant role by analyzing tourist behavior, accommodation and destination facilities (Singh et al. \u003cspan citationid=\"CR93\" class=\"CitationRef\"\u003e2023\u003c/span\u003e; Nath et al. \u003cspan citationid=\"CR68\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). ML models are used to look at tourist database through pattern recognition, data driven insights, adaptability and flexibility, improved accuracy, scalability and automation (Danish et al. 2023; Chien et al. \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). In this research, we have applied several machine models such as, i) Conditional Inference Tree (Kuhn and Johnson 2013) ii) Bagged CART (Hamze-Ziabari and Bakhshpoori \u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e2018\u003c/span\u003e) iii) Random Forest Model (Gevrey et al. \u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e2003\u003c/span\u003e) iv) Random Forest with Conditional Inference Tree (Quinlan \u003cspan citationid=\"CR75\" class=\"CitationRef\"\u003e1992\u003c/span\u003e), and v) Gradient Boosting Models (Kuhn and Johnson 2013). Further, combining the forecasts of various boot strapping base models such as decision trees, support vector machines, bagged CART, and Gradient Boosting models, an ensemble model was prepared, which eliminate the individual model's shortcomings and benefit from their advantages. Moreover, for the spatial assessments, the ML results can be linked with GIS. This makes it possible to represent spatial data in visual form (maps), in order to identify where TPZs can be developed. This integration of the data makes results easier to understand and provides users with added spatial context during decision-making.\u003c/p\u003e \u003cp\u003eTherefore, considering the perspectives outlined above, the research objective is to predict the TPZ using several machine learning, and ensemble models and also making an inter model comparison. The remaining manuscript was structured as follows:\u003c/p\u003e \u003cp\u003e \u003cul\u003e \u003cli\u003e \u003cp\u003eThe second section was marked with the introduction of the study area.\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003eThe third section was identified with the materials and method section.\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003eThe fourth section of the manuscript identifies the \u0026lsquo;Result\u0026rsquo; section of the manuscript.\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003eThe fifth section highlights the \u0026lsquo;Discussion\u0026rsquo; section and\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003eThe last section highlights the conclusion section in the manuscript.\u003c/p\u003e \u003c/li\u003e \u003c/ul\u003e \u003c/p\u003e"},{"header":"2. Materials and methods","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003e2.1 Study area\u003c/h2\u003e \u003cp\u003eThe Assam, the largest state in terms of people and second-largest in terms of territory, is bordered to the north by Bhutan and the Arunachal Pradesh, Nagaland, Arunachal Pradesh, and Manipur to the east, Bangladesh and Meghalaya and Tripura Mizoram and West Bengal to the south. According to Assam State Portal (assam.gov.in). Initially the Assam state had 27 districts however in 2016 about 5 districts were added however for this research and simplification total 27 districts were considered. The fact that Assam is home to three of India's six physiographic divisions\u0026mdash;the Northern Himalayas (Eastern Hills), Northern Plains (Brahmaputra plain), and Deccan Plateau is an important geographical feature (Karbi Anglong). Assam often experiences a \"Tropical Monsoon Rainforest Climate,\" with high humidity and significant precipitation. Warm summers and mild winters make for a temperate climate that residents may take advantage of all year round. The seasons of spring (March\u0026ndash;April) and fall (September\u0026ndash;October) are often pleasant with mild temperatures and rains. According to the Census of India (2011) (\u0026ldquo;Census Tables | Government of India\u0026rdquo;), Assam has a total population of 3.12 Cr. Assam's population makes up 2.58 percent of all Indians in 2011. Assam has a total population of 31,205,576 people, with 15,939,443 men and 15,266,133 women. Assam has a total size of 78,438 square kilometers. As a result, Assam has a population density that is greater than the national average.\u003c/p\u003e \u003cp\u003eAssam is extremely important for drawing tourists because of its magnificent mountains, biodiversity, plenty of foliage, ethnic diversity (fairs and festivals), and emission zone (Huismann, \u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e2014\u003c/span\u003e). Despite the country's popularity for tourists visiting, there are still many opportunities to explore its scenic beauty, grasslands (Bugyals), caves, bird watching sites, camping grounds, parks, and wildlife sanctuaries as well as its skiing areas, river valleys, passes, glaciers, mountain peaks, trekking trails, and river rafting locations (Choden et al. \u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e2018\u003c/span\u003e). The Himalaya region's mountains actually give people, especially tha chance toescape the pre-monsoon heat because of the good weather and picturesque scenery (Huismann \u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e2014\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eTill date, to the best of our knowledge, this research is the first to highlight on the tourism potentialities of the entire Assam. Previous research works highlighted the tourism potentials only in certain pockets of the state. For example, the nature-based tourism potentials of the Tinsukia District of the Upper Assam were divulged by Bordoloi and Agarwal (\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e2015\u003c/span\u003e). The tourism potentials of the Karabi Anglong autonomous council districts were estimated by Ronghang and Sen (2022). But the previous research activities completely neglected the spatial assessments, model building and its\u0026rsquo; validation. This research emphasized the tourism potentials of the Assam state using machine learning algorithms, making it an appealing reading for academicians and tourism practitioners. Therefore, this research is novel and has its\u0026rsquo; implications on assessments of tourism potentiality.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec4\" class=\"Section2\"\u003e \u003ch2\u003e2.2 First Step-Preparation of Tourism Inventory Database (TID)\u003c/h2\u003e \u003cp\u003eThe first step of the methodology (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e) was started by structuring the Tourism Inventory Database (TID). TID was important to understand assess the relationship that exists between the distribution of tourism potentiality and pre-depositing factor of tourism. The tourist spots of the present study area are characterized by archaeological, natural attraction, gift shops, guest houses, hotels, memorials, monuments, museums, parks, restaurants, churches, masjid, Hindu temples, arts centre, observation towers, and stadiums. Prominent archaeological spots which were identified are Namdang Xilor Xaaku and Sri Surya Pahar. Some prominent attractions of the study region are Sualkuchi, Gateway of Assam, Sukreshwar Ghat, Kamalabari, Bhogpur Satra, Barali Range mountain etc. Prominent gift shops are Arfed, Purbashree and The Eco Hub. The Guest Houses found in the state are Niribili Guest House, Subansiri Lodge, green Red Resort, Kaziranga Guest House, Drongo Guest House etc. Some hotels and restaurants are Mayuri, Prathma, Aroma Residency, Vishal, Heritage, Appayan Restaurant, TerraMaya, Di Cafeteria, Four Season\u0026rsquo;s Restaurant etc. Children\u0026rsquo;s Memorial and Children Educational and Career Development Centre are two prominent memorials in this state. Assam State Museum, Guwahati Planetarium, State Art Gallery, and Tai Museum are very renowned in the study area. Padum Pukhuri, Trimurti Udyan, Marut Kandan, Gandhi Bagh, Kaziranga National Orchid and Biodiversity, are the few examples of parks identified in the study region. Some Monasteries in the study area are Tinsukia Buddha Vihar, Dibong Buddhist Monastery, Kumchai Kong Monastery, Namphakey Buddhist Monastery, Powai Mukh Buddhist Monastery etc. This state is also enriched with Hindu temples (i.e., Kamalabari Temple, Ghanta Temple, Shivdham Temple, Krishna Dham Temple, Nepali Temple, Maa Kamakhya Temple etc.) and churches (i.e., Dibrugarh Charch, Beno Nomaya, Northern Evangelical Lutheran Church, Baptist Church, Christ Church etc.). This section also very enriched with Pure Muslim culture and traditions. Mosques (i.e., Duliajan Mosque, Fulertal Mukam, Bongaigaon Mosque etc.) bear witness to Islamic culture. Bihu Toli is a prominent Arts Centre in the Assam. Kachujan Stadium, Maligaon Railway Stadium, Nurul Amin Stadium, Rajiv Gandhi Indoor Stadium etc. are some of the important stadiums identified in the region. This region bears a perfect tourist viewshed with the help of several observational tower such as, Charboli, Fort, Kuri Bill Camp, Latajhar etc. Figure\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e (a-m) highlights some of the tourist spots identified during the field conducted in between 19.04.2021 to 20.06.2022 in three phases. Different tourist spots were identified (latitude and longitude were collected using GPS) by the non-participant observation technique (Banik and Mukhopadhyay \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). Further the official government websites were also consulted (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://tourism.assam.gov.in/portlets\u003c/span\u003e\u003cspan address=\"https://tourism.assam.gov.in/portlets\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec5\" class=\"Section2\"\u003e \u003ch2\u003e3.2 Second Step- Choosing tourism potentiality causative factors and multicollinearity\u003c/h2\u003e \u003cp\u003eTourism potentiality is a multidimensional concept in which several physical and socio-ecological variables are linked into it (Raha and Gayen \u003cspan citationid=\"CR77\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). The second step of the methodology (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e) was started by choosing the tourism potentiality causative factors and measurement of multicollinearity. Here, 11 criteria were choosed by the recommendation of 5 expert panel (Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e) and in-depth literature reviews (i.e., Raha and Gayen \u003cspan citationid=\"CR76\" class=\"CitationRef\"\u003e2023\u003c/span\u003e; Sahani \u003cspan citationid=\"CR86\" class=\"CitationRef\"\u003e2020\u003c/span\u003e; Trukachev 2015; Schultze et al. \u003cspan citationid=\"CR90\" class=\"CitationRef\"\u003e2014\u003c/span\u003e; Gourabi and Rad \u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e2013\u003c/span\u003e; Hoang et al. \u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e2018\u003c/span\u003e). The specialists were those with a minimum of five years of expertise in the travel and tourism industry. In a separate consent form, the experts, who were consulted gave their permission for the results to be used purely for academic purposes without revealing their original identity. Criteria in this research include relief, aspect, viewshed, forest area, wetland, coefficient of variation of rainfall, the reserved forest, population density, population growth rate, literacy rate and road/railway density. The relief, aspect, forest area, wetland, coefficient of variation of rainfall and reserved forest create initial base of the tourism activities (Yuxi and Linsberg 2020). A regional population structure is highlighted through the societal indicators.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eConsidered thematic layers and their sources of data\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"3\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCriteria\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eSources of data\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eDirectionality of influence\u003c/p\u003e \u003cp\u003eDescriptions\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eRelief (RL)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eSRTM DEM, spatial Resolution 90 m\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eThe moderate to low relief and aspect helps to vibrate the tourism potentialities positively (Codrea et al. \u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e2022\u003c/span\u003e).\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAspect (AS)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eSRTM DEM, spatial Resolution 90 m\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eViewshed (VS)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eSRTM DEM, spatial Resolution 90 m\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eIf the viewshed increases, tourism potentialities also increase and vice-versa (Sahani \u003cspan citationid=\"CR86\" class=\"CitationRef\"\u003e2020\u003c/span\u003e).\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eForest area (FA)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eIndia State Forest Report 2019\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\" morerows=\"3\" rowspan=\"4\"\u003e \u003cp\u003eForest area, wetland and rainfall variation and reserved forest positively enhance the tourism potentialities (Deribew et al. \u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e2022\u003c/span\u003e).\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eWetland (WL)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eAssam Project On Forest and Biodiversity Conservation (Apfbc) Society\u003c/p\u003e \u003cp\u003e\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://apfbcs.nic.in/apfbcs/wetland/annexure2.pdf\u003c/span\u003e\u003cspan address=\"https://apfbcs.nic.in/apfbcs/wetland/annexure2.pdf\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCV of Rainfall (CVR)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eClimate Research and Service Unit, Pune, (2020)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eReserve Forest (RF)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eAssam Project On Forest and Biodiversity Conservation (Apfbc) Society\u003c/p\u003e \u003cp\u003e\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://apfbcs.nic.in/apfbcs/wetland/annexure2.pdf\u003c/span\u003e\u003cspan address=\"https://apfbcs.nic.in/apfbcs/wetland/annexure2.pdf\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePopulation Density (PD)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eStatistical Hand Book of Assam, 2016\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eThe population density and growth rate indicate the balanced population structure therefore, it motivates the positive outcomes of tourism potentiality (Fakfare et al. \u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e2020\u003c/span\u003e)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePopulation Growth Rate (PGR)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eStatistical Hand Book of Assam, 2016\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLiteracy Rate (LR)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eStatistical Hand Book of Assam, 2016\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eThe literacy rate and the RRD both positively motivates the potential value for the tourism.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eRoad \u0026amp; Railway Density (RRD)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eOpen Street Map\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eThe sources of data were highlighted in the Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e. The relief, aspect and viewpoint were prepared from the Digital Elevation Model prepared by the Shuttle Radar Topographic Mission (SRTM) (Resolution is 1 ARC, pixel depth was 16 bits, published in 2014). The SRTM DEM was downloaded from the NASA Earth Explorer(\u0026ldquo;EarthExplorer,\u0026rdquo;). The information about the forest area was obtained from the India State Forest Report (2019). Information about the wetland and reserved forest were acquired from the Assam Project On Forest and Biodiversity Conservation (Apfbc) Society (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://apfbcs.nic.in/apfbcs/wetland/annexure2.pdf\u003c/span\u003e\u003cspan address=\"https://apfbcs.nic.in/apfbcs/wetland/annexure2.pdf\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e). The Coefficient of Variation of Rainfall data (2020) was obtained from the Climate Research and Service Unit, Pune, (2020). The Population Density, Population Growth Rate and Literacy Rate were obtained from the Statistical Handbook of Assam (2016).\u003c/p\u003e \u003cp\u003eThe multicollinearity is an essential prechecking before developing a particular model (Memon et al. \u003cspan citationid=\"CR60\" class=\"CitationRef\"\u003e2019\u003c/span\u003e). A high degree of correlation between two or more dependent variables; creates disturbance in the predictability of the model. In this research, the tolerance and Variance Intrusion Factor (VIF) methods were utilized to detect it. The formula was as follows (Eq.\u0026nbsp;\u003cspan refid=\"Equ1\" class=\"InternalRef\"\u003e1\u003c/span\u003e and Eq.\u0026nbsp;\u003cspan refid=\"Equ2\" class=\"InternalRef\"\u003e2\u003c/span\u003e):\u003cdiv id=\"Equ1\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equ1\" name=\"EquationSource\"\u003e\n$$Tolerence of i-th predictor variable \\left({T}_{i}\\right)=1-{R}_{i}^{2}$$\u003c/div\u003e\u003cdiv class=\"EquationNumber\"\u003e1\u003c/div\u003e\u003c/div\u003e\u003cdiv id=\"Equ2\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equ2\" name=\"EquationSource\"\u003e\n$$VIF of i-th predictor variable \\left({T}_{i}\\right)=\\frac{1}{1-{R}_{i}^{2}}$$\u003c/div\u003e\u003cdiv class=\"EquationNumber\"\u003e2\u003c/div\u003e\u003c/div\u003e\u003c/p\u003e \u003cp\u003eWhere, \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\({R}_{i}^{2}\\)\u003c/span\u003e\u003c/span\u003e depicts the coefficient of determination in the regression equation.\u003c/p\u003e \u003cp\u003eFor the independency, the tolerance level should be more than 0.10 and the VIF should be less than 10.0.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec6\" class=\"Section2\"\u003e \u003ch2\u003e3.3 Third Step-Spatial modelling of TPZ\u003c/h2\u003e \u003cp\u003eIn the third step (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e), TPZs were predicted using the single classification tree model (ctree model), Bagged CART (treebag model), Random Forest model (rf model), Random Forest with bagging ensemble algorithms utilizing conditional inference tree (cforest model) and Gradient Boosting (gbm model) models. Further, combining the ctree, treebag, rf, cforest and gbm models a new ensemble model was prepared. For this research, we have set the scenario as the 10-fold cross validation with repeats in train control. Here, the objective with cross-validated data set was to optimize and determine the size of the tree by tuning the complexity parameters.\u003c/p\u003e \u003cdiv id=\"Sec7\" class=\"Section3\"\u003e \u003ch2\u003e3.3.1 Conditional Inference Tree (ctree model)\u003c/h2\u003e \u003cp\u003eThe ctree model is a non-parametric class of regression trees embedding tree structured regression models into well-defined theory of conditional inference procedures. It is applicable to all kind of regression problems including multiple response scales of covariates (Hothorn et al. \u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e2015\u003c/span\u003e; Fu \u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e2017\u003c/span\u003e). The response \u003cem\u003eY\u003c/em\u003e given the status of m covariates by means of tree structured recursive partitioning. The m dimensional covariate vector \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(X=({X}_{1},\\dots \\dots ..,{X}_{m})\\)\u003c/span\u003e\u003c/span\u003e was taken from the sample space \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\aleph ={\\aleph }_{1\\dots \\times \\dots \\dots \\dots \\dots ..}\\times {\\aleph }_{m.}\\)\u003c/span\u003e\u003c/span\u003eHere both response and covariates could be measured in the arbitrary scales. The conditional function distribution D(Y/X) of the response Y given the covariates X depends on the function \u003cem\u003ef\u003c/em\u003e of covariates (Eq.\u0026nbsp;\u003cspan refid=\"Equ3\" class=\"InternalRef\"\u003e3\u003c/span\u003e) (Hothorn et al. \u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e2015\u003c/span\u003e)\u003cdiv id=\"Equ3\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equ3\" name=\"EquationSource\"\u003e\n$$D\\left(Y|X\\right)=D\\left(Y|{X}_{1},\\dots \\dots \\dots \\dots \\dots .{X}_{m}\\right)=D\\left(Y\\right|f\\left({X}_{1,\\dots \\dots \\dots \\dots ..}{X}_{m}\\right)$$\u003c/div\u003e\u003cdiv class=\"EquationNumber\"\u003e3\u003c/div\u003e\u003c/div\u003e\u003c/p\u003e \u003cp\u003eWhere, we restrict ourselves to portion based regression relationships i.e., r disjoint cells, \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\({B}_{1,\\dots \\dots ..}{B}_{r, }\\)\u003c/span\u003e\u003c/span\u003epartitioning covariate space i.e., \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\aleph ={U}_{K=1}^{r}{B}_{K.}\\)\u003c/span\u003e\u003c/span\u003e\u003c/p\u003e \u003cp\u003eThe regression relationship should be fitted on a learning sample \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\({L}_{n}\\)\u003c/span\u003e\u003c/span\u003e i.e, learning samples with n independent observations possibly with same coordinates \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\({X}_{ij}\\)\u003c/span\u003e\u003c/span\u003e missing (Eq.\u0026nbsp;\u003cspan refid=\"Equ4\" class=\"InternalRef\"\u003e4\u003c/span\u003e)\u003cdiv id=\"Equ4\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equ4\" name=\"EquationSource\"\u003e\n$${L}_{n}=\\{\\left({Y}_{i,\\dots ..}{X}_{1i,\\dots \\dots \\dots \\dots .,}{X}_{mi}\\right);i=1,\\dots \\dots .,n\\}$$\u003c/div\u003e\u003cdiv class=\"EquationNumber\"\u003e4\u003c/div\u003e\u003c/div\u003e\u003c/p\u003e \u003cp\u003eA genetic algorithm for the recursive binary partitioning for a given learning sample \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\({L}_{n}\\)\u003c/span\u003e\u003c/span\u003e can be formulated using non-negative integer valued case weights \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(w=\\left({w}_{1},\\dots \\dots \\dots \\dots ,{w}_{n}\\right).\\)\u003c/span\u003e\u003c/span\u003e Each node of the tree is attached with non-zero or zero weights as the case may be.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec8\" class=\"Section3\"\u003e \u003ch2\u003e3.3.2 Bagged CART model (treebag model)\u003c/h2\u003e \u003cp\u003eBagging, which stands for Bootstrap Aggregating, is a technique used to enhance the stability and accuracy of machine learning algorithms, in particular for decision trees, such as CART (treebag model) (Vrontos et al. \u003cspan citationid=\"CR105\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). This is how it worked within this research: a) Multiple bootstrap samples were created in each iteration, and each sample was essentially a random sample with replacement and is of the same size as the original dataset. B) For CART, a decision tree model was trained for each of the hundreds of bootstrap samples, which was slightly different from each other since each training sample was different in the first place. After all decision trees were trained, the predictions for the new data was made by aggregating all the predictions made by all these individual trees. For the classification tasks in this research, it was done by majority voting \u0026ndash; the class out of all, that has the most votes. It helps to reduce the overfitting because bagging trains many models on slightly different datasets, and aggregates all their predictions together. So, it lowers the variance of the whole final model (Choi and Hur \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2020\u003c/span\u003e; Gonz\u0026aacute;lez et al. \u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e2020\u003c/span\u003e).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec9\" class=\"Section3\"\u003e \u003ch2\u003e3.3.3 Random forest model (rf model)\u003c/h2\u003e \u003cp\u003eOne of the most popular supervised learning models is Random Forest (rf), which was used as a classification model in this research. Breiman (\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e2001\u003c/span\u003e) first developed rf model algorithm with the help of decision trees (Zang et al. 2017; Gayen et al. \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e2019\u003c/span\u003e). Its\u0026rsquo; basis is exactly the method of ensemble learning, in which a user has the opportunity to \u0026ldquo;tie\u0026rdquo; multiple classifiers together to address challenging tasks and enhance the performance of the model. One of the assets of Random Forest is essentially a decision tree. The final decision is made by result polling from different trees and selecting the most popular one. Most of the outcomes of each tree generate the end output. Random forest model provides more accuracy than the other model and also provides a clear and separate distribution plot of features in each class (Couronn\u0026eacute;, et al. \u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e2018\u003c/span\u003e; Fox et al. \u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e2017\u003c/span\u003e). It can handle the missing data effectively; the developed model can be saved for future use with the new data. In this research following steps were followed to build a Random Forest model:\u003c/p\u003e \u003cp\u003e \u003cul\u003e \u003cli\u003e \u003cp\u003eFirst of all, \u0026lsquo;K\u0026rsquo; features from total m features were randomly selected, where k\u0026thinsp;\u0026lt;\u0026thinsp;m\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003eNode \u0026lsquo;d\u0026rsquo; was calculated among \u0026lsquo;K\u0026rsquo; features.\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003eSplitted the node into several daughter nodes using the best split method.\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003eRepeated the previous steps until reaching the \u0026lsquo;I\u0026rsquo; number of nodes.\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003eBuilt a forest by repeating all steps for \u0026lsquo;n\u0026rsquo; number of times to create \u0026lsquo;n\u0026rsquo; number of trees.\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003eThe mean squared error of each decision tree with their OOB samples (E\u003csub\u003eOOB\u003c/sub\u003e) is used to calculate the learning error.\u003c/p\u003e \u003c/li\u003e \u003c/ul\u003e \u003c/p\u003e \u003cp\u003eThe pros of this approach are \u0026ndash; i) that it can deal with very voluminous data, of sufficiently large dimensions and escape the risk of overfitting (Naghibi et al. \u003cspan citationid=\"CR67\" class=\"CitationRef\"\u003e2017\u003c/span\u003e); ii) this does not imply additional assumptions about the factors to be manipulated and the result (Youssef et al. \u003cspan citationid=\"CR102\" class=\"CitationRef\"\u003e2016\u003c/span\u003e); iii) the analyst needs a working dataset, as there is no transformation and scalability prior to the procedure (Gayen et al. \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e2019\u003c/span\u003e) and iv) the model is easily applicable to any regional scale (Pourghasemi and Rahmati 2019).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec10\" class=\"Section3\"\u003e \u003ch2\u003e3.3.4 Conditional Inference Random Forest (cforest model)\u003c/h2\u003e \u003cp\u003eThe cforest (Conditional Inference Random Forest) is a combination of random forest and bagging ensemble algorithm implementation applied in this research. The main advantage of the Cforest is that it uses conditional inference trees as its base learners (Naghibi et al. \u003cspan citationid=\"CR67\" class=\"CitationRef\"\u003e2017\u003c/span\u003e). Cforest also uses OOB data, although that means more info and better accuracy, it is also slower and can handle less data for the same memory (Thanh et al. \u003cspan citationid=\"CR99\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). The weighted average of the trees was used in this research to get the final ensemble. The main reason for the cforest\u0026rsquo;s more reliable predictions is the fact that it produces unbiased trees (Strobl et al. \u003cspan citationid=\"CR94\" class=\"CitationRef\"\u003e2007\u003c/span\u003e; Mogensen et al. \u003cspan citationid=\"CR61\" class=\"CitationRef\"\u003e2012\u003c/span\u003e). The cforest is always better when the model has computational resources.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec11\" class=\"Section3\"\u003e \u003ch2\u003e3.3.5 Gradient Boosting model (gbm model)\u003c/h2\u003e \u003cp\u003eGradient Boosting (gbm model) offer a powerful technique for tackling regression and classification problems, leveraging ensemble learning by strategically combining multiple weak algorithms into a robust solution (Islam et al. \u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). This method was also applied in this research. Typically employing decision trees as its basic components, this approach incrementally builds a predictive model through an iterative process, gradually boosting accuracy as each new element is incorporated into the evolving whole (Zhang et al. 2019; Sachdeva and Kumar \u003cspan citationid=\"CR85\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). The initial phase establishes a set of basic learners, which are shallow decision trees being a common choice due to their simplicity - to serve as foundations. Then, through gradient boosting, subsequent rounds fit new weak learners to the residual errors of their predecessors, reducing discrepancies between actual and predicted outcomes turn by turn. By minimizing collective mistake residuals, each additional tree tuned to those of earlier stages, the ensemble members are combined in an optimized manner that strengthens overall predictive capabilities with every included model. This model uses gradient descent optimization in order to minimize the loss function (Ridgeway \u003cspan citationid=\"CR79\" class=\"CitationRef\"\u003e2007\u003c/span\u003e). At every iteration, it computes the gradient of the loss function with respect to the predictions of the ensemble and then it updates the predictions in the direction which minimizes the loss (Lu et al. \u003cspan citationid=\"CR56\" class=\"CitationRef\"\u003e2020\u003c/span\u003e;. In order to fight overfitting, gbm model also uses shrinkage or learning rate. Instead of using additions of smaller number of trees, a contribution of a less than one scaled of each tree are added to the ensemble. Smaller contributions will make the optimization faster and more conservative which makes it require a higher number of iterations but will protect the model against overfitting. It can also use some kind of regularization; it can be either tree pruning which removes some of the splits producing no positive results or it can limit the maximum possible depth of the trees.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec12\" class=\"Section3\"\u003e \u003ch2\u003e3.3.6 Ensemble model\u003c/h2\u003e \u003cp\u003eBy combining the above models an ensemble model was prepared and applied for the prediction of TPZ. In this research it is regarded as \u0026lsquo;\u003cem\u003eTPZ Ensemble Model\u0026rsquo;\u003c/em\u003e. An ensemble model is a machine learning method involving combining single models for producing a more precise prediction far more robust than any one model individually (Ganaie et al. \u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e2022\u003c/span\u003e; Mohammed et al. 2023). The rationale for this ensemble model is consistent with the so-called \u0026ldquo;wisdom of the crowd,\u0026rdquo; meaning that the average belief of numerous models (here, ctree, treebag, rf, cforest and gbm models) is more efficient and reliable than the view of any particular model. In this research, stacking, or stacked generalization, was used to build the \u0026lsquo;TPZ Ensemble model\u0026rsquo;. Stacking is a very popular method, which was used in this research to prepare the ensemble model. Initially we trained base models such as, ctree, treebag, rf, cforest and gbm models. Next, instead of simply averaging or voting; another meta model or blender was introduced here. After that, the meta model was trained to learn the best way to combine the predictions of the base models to make the final prediction. During the prediction phase, the base models make their predictions on new data, and then the meta-model combines these predictions to give the final output (Polikar \u003cspan citationid=\"CR71\" class=\"CitationRef\"\u003e2006\u003c/span\u003e; Optiz and Maclin 1999). Different base models capture different aspects or patterns within the data. By using all of them, we combined to the strengths on these models, and negate their weaknesses. It introduces diversity among the models, which helps it be less vulnerable to overfitting (Gomes et al. \u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e2017\u003c/span\u003e). If one model overfits to some patterns within the data, others might capture different patterns or generalize better towards unseen data (Dong et al. \u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e2020\u003c/span\u003e; Opitz and Maclin \u003cspan citationid=\"CR69\" class=\"CitationRef\"\u003e1999\u003c/span\u003e; Polikar \u003cspan citationid=\"CR71\" class=\"CitationRef\"\u003e2006\u003c/span\u003e). Ensemble methods are more robust to noise and outliers (Mienye and Sun \u003cspan citationid=\"CR64\" class=\"CitationRef\"\u003e2022\u003c/span\u003e; Rokach \u003cspan citationid=\"CR81\" class=\"CitationRef\"\u003e2010\u003c/span\u003e). For instance, if a particular model\u0026rsquo;s prediction is unreliable when the input data is outside the training data, ensembles can help improve this prediction using other model\u0026rsquo;s reliable predictions made based on that input data (Blockeel \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e2011\u003c/span\u003e; Rincy and Gupta \u003cspan citationid=\"CR80\" class=\"CitationRef\"\u003e2020\u003c/span\u003e; Ganaie et al. \u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). In this research, all models were computed in the RStudio Version 2023.12.1 with Intel(R) Core (TM) i5-9300H CPU @ 2.40GHz 2.40 GHz processor with 8GB Ram and 64-bit operating system,\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv id=\"Sec13\" class=\"Section2\"\u003e \u003ch2\u003e3.5 Fourth Step- accuracy assessments\u003c/h2\u003e \u003cp\u003eOne of the most crucial aspects of model building is the evaluation of the output models' precision (Das \u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). Here, the TPZ was validated using the Kappa Coefficient, Accuracy and AUC-ROC curve. The ROC-AUC shows how specificity and sensitivity are traded off. The ROC is a two-dimensional graph in which x-axis depicts the specificity and the y axis depicts the sensitivity. The euqations 5, 6 and 7 illustrated the attributes of x and y axis, where, the TN represents true negative, FP represents false positive, TP represents true positive, and FN represents false negative (Eq.\u0026nbsp;\u003cspan refid=\"Equ5\" class=\"InternalRef\"\u003e5\u003c/span\u003e, Eq.\u0026nbsp;\u003cspan refid=\"Equ6\" class=\"InternalRef\"\u003e6\u003c/span\u003e and Eq.\u0026nbsp;\u003cspan refid=\"Equ7\" class=\"InternalRef\"\u003e7\u003c/span\u003e) (Roodposhti et al. \u003cspan citationid=\"CR83\" class=\"CitationRef\"\u003e2017\u003c/span\u003e):\u003cdiv id=\"Equ5\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equ5\" name=\"EquationSource\"\u003e\n$$x=specificity=\\left[\\frac{TN}{(TN+FP)}\\right]$$\u003c/div\u003e\u003cdiv class=\"EquationNumber\"\u003e5\u003c/div\u003e\u003c/div\u003e\u003cdiv id=\"Equ6\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equ6\" name=\"EquationSource\"\u003e\n$$y=sensitivity=\\left[\\frac{TP}{(TP+FN)}\\right]$$\u003c/div\u003e\u003cdiv class=\"EquationNumber\"\u003e6\u003c/div\u003e\u003c/div\u003e\u003cdiv id=\"Equ7\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equ7\" name=\"EquationSource\"\u003e\n$$Accuracy=\\left[\\frac{TP+TN}{(TP+TN+FP+FN)}\\right]$$\u003c/div\u003e\u003cdiv class=\"EquationNumber\"\u003e7\u003c/div\u003e\u003c/div\u003e\u003c/p\u003e \u003cp\u003eThe Kappa coefficient was defined as (Eq.\u0026nbsp;\u003cspan refid=\"Equ8\" class=\"InternalRef\"\u003e8\u003c/span\u003e)\u003cdiv id=\"Equ8\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equ8\" name=\"EquationSource\"\u003e\n$$Kappa=\\left[\\frac{{P}_{0}-{P}_{est}}{1-{P}_{est}}\\right]$$\u003c/div\u003e\u003cdiv class=\"EquationNumber\"\u003e8\u003c/div\u003e\u003c/div\u003e\u003c/p\u003e \u003cp\u003eWhere, \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\({P}_{0}\\)\u003c/span\u003e\u003c/span\u003e is defined as the observed agreement; and \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\({P}_{est}\\)\u003c/span\u003e\u003c/span\u003e is the expected agreement.\u003c/p\u003e \u003cp\u003eThe performance of the model is quantitatively depicted by the area under the ROC curve (AUC) (Tang et al. \u003cspan citationid=\"CR97\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). A standard scale of AUC is i)\u0026thinsp;\u0026ge;\u0026thinsp;0.9 denotes excellent ii) 0.8 to 0.9 denotes accepted iii) 0.7 to upto 0.8 denotes good or satisfactory, iv) 0.5 to 0.7 is considerable and v) less than 0.5 is rejected (Trabelsi et al. \u003cspan citationid=\"CR100\" class=\"CitationRef\"\u003e2023\u003c/span\u003e; Mitra et al. \u003cspan citationid=\"CR65\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). It is recommended that the machine learning models should be judged based on the validation or test dataset (Vabalas et al. \u003cspan citationid=\"CR103\" class=\"CitationRef\"\u003e2019\u003c/span\u003e). Because the validation sets are commonly used for hyperparameter tuning where different hyperparameters configurations of the model are tested to find the best-performing one. This ensures that the model\u0026rsquo;s performance is optimized for the specific dataset while still keeping its ability to generalize.\u003c/p\u003e \u003c/div\u003e"},{"header":"4. Results and discussion","content":"\u003cdiv id=\"Sec15\" class=\"Section2\"\u003e\n\u003ch2\u003e4.1 Analysis of multicollinearity\u003c/h2\u003e\n\u003cp\u003eFor all criteria, the tolerance level fluctuated from .322 (for population density) to .977 (for aspect) and the VIF varied from 1.217 (for reserved forest) to 2.839 (for forest area) (Table\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e2\u003c/span\u003e). The highest VIF assures the lowest tolerance level. Here, all VIF values are less than 10 and all tolerance level values are less than 1.0. Therefore, this research shows that multicollinearity is absent in the eleven pre-processing factors being investigated.\u003c/p\u003e\n\u003cdiv class=\"gridtable\"\u003e\n\u003ctable id=\"Tab2\" border=\"1\"\u003e\u003ccaption\u003e\n\u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e\n\u003cdiv class=\"CaptionContent\"\u003e\n\u003cp\u003eTolerance level and Variance Intrusion Factor for different indicators\u003c/p\u003e\n\u003c/div\u003e\n\u003c/caption\u003e\n\u003cthead\u003e\n\u003ctr\u003e\n\u003cth align=\"left\"\u003e\n\u003cp\u003eCriteria\u003c/p\u003e\n\u003c/th\u003e\n\u003cth align=\"left\"\u003e\n\u003cp\u003eTolerance\u003c/p\u003e\n\u003c/th\u003e\n\u003cth align=\"left\"\u003e\n\u003cp\u003eVIF\u003c/p\u003e\n\u003c/th\u003e\n\u003c/tr\u003e\n\u003c/thead\u003e\n\u003ctbody\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e\u003cstrong\u003eRL\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e.491\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e2.036\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e\u003cstrong\u003eAS\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e.977\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e1.023\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e\u003cstrong\u003eCVR\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e.495\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e2.022\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e\u003cstrong\u003eFA\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e.352\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e2.839\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e\u003cstrong\u003eVS\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e.595\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e1.681\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e\u003cstrong\u003eRF\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e.822\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e1.217\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e\u003cstrong\u003eWL\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e.756\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e1.323\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e\u003cstrong\u003ePGR\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e.496\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e2.017\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e\u003cstrong\u003eLR\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e.458\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e2.182\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e\u003cstrong\u003ePD\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e.322\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e3.102\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e\u003cstrong\u003eRRD\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e.814\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e1.229\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003c/tbody\u003e\n\u003c/table\u003e\n\u003c/div\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec16\" class=\"Section2\"\u003e\n\u003ch2\u003e4.2 Spatial interrelationship between tourism location and causative factors\u003c/h2\u003e\n\u003cp\u003eDistribution of tourism locations and causative factors were illustrated in this research using Analytic Hierarchy Process (AHP) model. The AHP is an unbiased MCDM method for choosing the best option from a large pool of alternatives (Munier and Hontoria \u003cspan class=\"CitationRef\"\u003e2021\u003c/span\u003e; Senapati and Das \u003cspan class=\"CitationRef\"\u003e2021\u003c/span\u003e). The AHP method was invented by Saaty, (\u003cspan class=\"CitationRef\"\u003e1980\u003c/span\u003e,\u003cspan class=\"CitationRef\"\u003e1987\u003c/span\u003e), and it attracted a wide number of researchers for its\u0026rsquo; adaptability and usefulness. First, the pair-wise comparative matrix, which represents the relative priorities of each criterion, has an identical number of rows and columns was prepared. The 5-member expert panel chose the significance of several criteria in this case. The panel was created by including those experts and researchers, who had at least five years of experience in the field of travel and tourism. The experts who were consulted agreed, in a separate consent form, that the scores might be used only for academic purposes, without disclosing their identities. The preference of each criterion was estimated using a relative dominance scale of 1 to 9 (Saaty \u003cspan class=\"CitationRef\"\u003e1980\u003c/span\u003e). Here, 1, 3, 5, 7 and 9 were marked as equal importance, moderate importance, strong importance, very strong importance and extreme or substantial importance. 2,4,6 and 8 represent intermediate values. The necessary condition of a considerable AHP matrix is that the consistency ratio or C.R. value should be less than 0.1,\u003c/p\u003e\n\u003cp\u003eThe relief of Assam varied from 1 meter to 1971meter (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e4\u003c/span\u003ea). The Western and the Eastern sections of the study area were with lower elevation (i.e.,1 meter to 300 meters). The southern sections are comparatively high (i.e., 300 meters to 1971 meters). The tourism potentiality decreases with the increasing relief value. As a result, priority rises as the class value of the relief decrease, and vice versa. 1 to 300-meter relief class had higher areal coverage (87.055% area). There also an inverse trend between the very high relief and tourist locations. Therefore, Tourism potentiality pixels are substantially overlapped within the moderate to low relief classes. The aspect fluctuated from \u0026minus;\u0026thinsp;1 to 359.458 (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e4\u003c/span\u003eb). Lower aspect is getting more attention in case of tourism and its potentiality. In this study, the AS was divided into four categories and importance increased as class value decreased. The first category (-1 to 89.211) was directed towards flat, northern, north-eastern and eastern direction. The Bramhaputra river passes in the middle portion of the Assam and along the river lowest aspect value was marked (1 to 89.211). The second category (89.212 to 179.424) was marked towards the eastern, south-eastern and southern directions. The third category (179.424 to 269.636) was directed towards the southern, south-west and western directions. The fourth category (269.637 to 359.848) was marked towards the western, northwestern and northern directions. As the distance increases from the river, the aspect value increases. Tourism potentiality pixels are largely overlapped in moderate to low aspect values (i.e., 194 pixels). The forest area fluctuated from 4.52\u0026ndash;86.07% within the study area. Larger forest area was marked in the districts of Karbi Anglong, West Karbi Anglong, Dima Hasao, Cachar, Hailakandi, and Karimganj. 30\u0026ndash;50% forest area was found in the Kakrajhar, Chirang, Golpara, Kamrup, Karimganj and Tinsukia districts. Remaining districts were marked with comparatively low forest area (i.e., 4.52\u0026ndash;30% area). The picturesque attractiveness is positively accelerated by the FA, which draws many adventure travelers (Karali et al. \u003cspan class=\"CitationRef\"\u003e2021\u003c/span\u003e). The FA was splitted into four groups for this study, and as the category value is higher, weightages likewise increase and vice versa (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e4\u003c/span\u003ec). Moderate to high forest cover attracts more tourism potentiality pixels (i.e., 162 pixels). The number of reserved forests in the Assam fluctuated from 0 to 29 (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e7\u003c/span\u003ed). Comparatively higher number of reserved forests (i.e., 16 to 29 numbers) were found in the West Karabi Arlong, Naogaon, Karbi Arlong, Tinsukia and Kamrup Metropolitan districts. Remaining districts were marked with lower number (0 to 15 numbers) of reserved forests. The number of protected forestry has a favorable impact on the tourism potential. Here, in this research, the RF was classified into four classes and the class value increases, weightages also increase and vice-versa (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e4\u003c/span\u003ed). Moderate and high classes of RF incur higher tourism potentiality pixels in the study area (i.e., 234 pixels). The number of wetlands in the study area varied from 0 to 1790 (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e5\u003c/span\u003ea). The higher number of wetlands (i.e., 251 to 1790 numbers) were identified in the districts of Kakrajhar, Karbi Arlong, Naogaon, West Karbi Arlong, Sontipur, Hailakandi, Karimganj and Tinsukia districts. The remaining portions were noticed with lower number of wetlands. The WL favorably vibrates the tourism potentiality. Here, the WL was broken down into four classes, and priority increases with the increasing class value. Tourism potentiality pixels also substantially merged under very high, high and moderate class values of WL (i.e., 228 pixels). The coefficient of rainfall variation (CVR) fluctuated from 86.401\u0026ndash;108.007% (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e5\u003c/span\u003eb). Most parts of the Assam experienced a large variation of rainfall. Comparatively higher amount of rainfall variation (92.889\u0026ndash;108.006%) was noticed in the Dhubri, Kokrajhar, Golpara, Bongajagaon, Barpeta, Chirang, Baksa, Nalbari, Kamrup, Karabi Arlong, Darang, Kamrup Metropolitan, Morigaon, Sontipur, Lakhimpur, Dhemji, Naogaon, Dima Hasaao, Cachar, Hailakandi and Kamrup districts. The remaining portions were noticed with lower amount of rainfall variation. The CVR positively enhances the potential for tourism (Giorgi and Lionello \u003cspan class=\"CitationRef\"\u003e2008\u003c/span\u003e). Four categories were used to classify the CV in this instance, and as the value of each category increases, so does priority, and vice versa. Tourism location pixels also substantially merged under very high, high and moderate class values of CVR (i.e., 228 pixels). The viewshed of the study area was marked by identifying 17 major hills (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e5\u003c/span\u003ec). Here, the viewshed was classified into 4 classes. The Western sections of the study area were more prominent. The VS positively enhances the potential for tourism (Giorgi and Lionello \u003cspan class=\"CitationRef\"\u003e2008\u003c/span\u003e; Ferguson et al. 2022). The VS was split into four categories in this instance, and priority increases as the class value increases. Tourism potentiality pixels are overlapped majorly over very high, high and moderate classes of VS (i.e. 273 pixels). The PGR of the research area varied from 5.210\u0026ndash;24.440% (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e6\u003c/span\u003ea). Comparatively high PGR value (i.e., 20 to 24%,) was found in the Dhubri, Bongaigaon, Golpara, Barpata, Darrang, Kamrup Metropolitan, Morigaon, Cacchar, Hailakandi and Karimganj districts. Remaining districts have 5.210\u0026ndash;20% growth rate. Upper Assam has comparatively low growth rate. The PGR in this instance was separated into four categories, and priority increases as the quantity of each class increases and vice versa. High, Very High and Moderate classes of PGR are identified with a higher number of pixels counts of tourism potentiality (i.e., 311 pixels).The PD of the study area varied from 44 to 1313 persons per square km. Comparatively high PD value (i.e., 701 to 1313 persons/sq.km,) was found in the Dhubri, Barpata, Nalbari, Kamrup Metropolitan and Naogaon districts. Remaining districts have 44 to 700 persons /square km (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e6\u003c/span\u003eb). High population density negatively affects the tourism potentiality. Therefore, the PD in this instance was divided into four separate classes, and importance increases as the class value decreases and vice versa. The substantial tourism potentiality pixels overlapped in the low and moderate PD classes (i.e., 315 pixels). The LR varied from 65.37\u0026ndash;88.71% within the study area of Assam. Comparatively low literacy rates were marked in the Dhubri, Chirang, Sontipur, Baksa, Golpara, Jorhat and Tinsukia districts. Higher literacy rates were identified in the Darrang, Kamrup Metropolitan, Morigaon, Golaghat, Jorhat, Sivsagar, Karbi Anglong, Hailakandi, and Karimganj districts (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e6\u003c/span\u003ec). As tourism is a tertiary activity, therefore, the LR would positively vibrate the tourism activities. Here, the LR was classified into four groups, and as the value of each class rises, so does priority, and vice versa. High, very high and moderate classes of LR combinedly achieve higher number of pixels for tourism potentiality (i.e., 310 pixels). Road and railway networks connect the tourism destinations quickly and smoothly (Bast et al. \u003cspan class=\"CitationRef\"\u003e2016\u003c/span\u003e). As a result, the RRD was divided into four categories. As the class value rises, so does priority, and vice versa. Moderate, high and very high classes of RRD combinedly overlapped with higher tourism potentiality pixel value (i.e., 249 pixels). The detailed areal coverage of different thematic layer, their classes and weightages were marked in Table\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e3\u003c/span\u003e.\u003c/p\u003e\n\u003ctable border=\"1\" width=\"728\"\u003e\u003ccaption\u003e\n\u003cp\u003eTable 3\u003c/p\u003e\n\u003cp\u003eRelationship between the variables with the help of the AHP method\u003c/p\u003e\n\u003c/caption\u003e\n\u003ctbody\u003e\n\u003ctr\u003e\n\u003ctd width=\"38\"\u003e\n\u003cp\u003e\u003cstrong\u003eIndicators\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u003cstrong\u003eClass\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd colspan=\"4\" width=\"265\"\u003e\n\u003cp\u003e\u003cstrong\u003eCategory\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u003cstrong\u003ePriority % (Weightage)\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"76\"\u003e\n\u003cp\u003e\u003cstrong\u003ePrincipal eigen value \u0026amp; Consistency Ratio (C.R.)\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"57\"\u003e\n\u003cp\u003e\u003cstrong\u003eNumber of Pixels in the domain\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"85\"\u003e\n\u003cp\u003e\u003cstrong\u003eNumber of tourism potentiality pixels in the domain\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"76\"\u003e\n\u003cp\u003e\u003cstrong\u003e% Area\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd rowspan=\"5\" width=\"38\"\u003e\n\u003cp\u003e\u003cstrong\u003eRL\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u003cstrong\u003e1.00 - 80.00\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e(Low)\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u003cstrong\u003e80.01 - 300.00\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e(Moderate)\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u003cstrong\u003e300.01 - 800.00\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e(High)\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u003cstrong\u003e800.01 - 1971.00\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e(Very High)\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"76\"\u003e\n\u003cp\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"57\"\u003e\n\u003cp\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"85\"\u003e\n\u003cp\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"76\"\u003e\n\u003cp\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u003cstrong\u003e1.00 - 80.00\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e1\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e3\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e4\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e5\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e(55.00%) 8\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd rowspan=\"4\" width=\"76\"\u003e\n\u003cp\u003ePrincipal eigen value =\u0026nbsp;4.057,\u003c/p\u003e\n\u003cp\u003eConsistency Ratio CR\u0026nbsp;=\u0026nbsp;2.1%\u0026nbsp;\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"57\"\u003e\n\u003cp\u003e3696933\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"85\"\u003e\n\u003cp\u003e140\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"76\"\u003e\n\u003cp\u003e39.234\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u003cstrong\u003e80.01 - 300.00\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e0.33\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e1\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e2\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e2\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e(21.40) % 6\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"57\"\u003e\n\u003cp\u003e44485020\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"85\"\u003e\n\u003cp\u003e178\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"76\"\u003e\n\u003cp\u003e47.281\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u003cstrong\u003e300.01 - 800.00\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e0.25\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e0.5\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e1\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e2\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e(14.20%) 4\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"57\"\u003e\n\u003cp\u003e9533666\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"85\"\u003e\n\u003cp\u003e37\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"76\"\u003e\n\u003cp\u003e10.249\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u003cstrong\u003e800.01 - 1971.00\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e0.2\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e0.5\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e0.5\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e1\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e(9.40%) 2\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"57\"\u003e\n\u003cp\u003e2508725\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"85\"\u003e\n\u003cp\u003e10\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"76\"\u003e\n\u003cp\u003e2.697\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd rowspan=\"5\" width=\"38\"\u003e\n\u003cp\u003e\u003cstrong\u003eAS\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u0026nbsp;\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u003cstrong\u003e-1 - 89.211\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e(Low)\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u003cstrong\u003e89.212 - 179.424\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e(Moderate)\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u003cstrong\u003e179.425 - 269.636\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e(High)\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u003cstrong\u003e269.637 - 359.848\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e(Very High)\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u0026nbsp;\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"76\"\u003e\n\u003cp\u003e\u0026nbsp;\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"57\"\u003e\n\u003cp\u003e\u0026nbsp;\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"85\"\u003e\n\u003cp\u003e\u0026nbsp;\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"76\"\u003e\n\u003cp\u003e\u0026nbsp;\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u003cstrong\u003e-1 - 89.211\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e1\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e2\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e5\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e7\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e(54.50%) 8\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd rowspan=\"4\" width=\"76\"\u003e\n\u003cp\u003ePrincipal eigen value =\u0026nbsp;4.072,\u003c/p\u003e\n\u003cp\u003eConsistency Ratio CR\u0026nbsp;=\u0026nbsp;2.6%\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"57\"\u003e\n\u003cp\u003e23798802\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"85\"\u003e\n\u003cp\u003e80\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"76\"\u003e\n\u003cp\u003e25.583\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u003cstrong\u003e89.212 - 179.424\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e0.5\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e1\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e2\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e3\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e(24.70%) 6\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"57\"\u003e\n\u003cp\u003e24023779\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"85\"\u003e\n\u003cp\u003e114\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"76\"\u003e\n\u003cp\u003e25.825\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u003cstrong\u003e179.425 - 269.636\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e0.2\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e0.5\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e1\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e3\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e(14.10%) 5\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"57\"\u003e\n\u003cp\u003e22381165\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"85\"\u003e\n\u003cp\u003e95\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"76\"\u003e\n\u003cp\u003e24.059\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u003cstrong\u003e269.637 - 359.848\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e0.14\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e0.33\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e0.33\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e1\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e(6.70%) 3\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"57\"\u003e\n\u003cp\u003e22820598\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"85\"\u003e\n\u003cp\u003e73\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"76\"\u003e\n\u003cp\u003e24.532\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd rowspan=\"5\" width=\"38\"\u003e\n\u003cp\u003e\u003cstrong\u003eFA\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u0026nbsp;\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u003cstrong\u003e50.001 - 86.070\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e(Very high)\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u003cstrong\u003e30.001- 50.00\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e(High)\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u003cstrong\u003e15.001 - 30.00\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e(Moderate)\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u003cstrong\u003e4.520 - 15.000\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e(Low)\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u0026nbsp;\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"76\"\u003e\n\u003cp\u003e\u0026nbsp;\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"57\"\u003e\n\u003cp\u003e\u0026nbsp;\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"85\"\u003e\n\u003cp\u003e\u0026nbsp;\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"76\"\u003e\n\u003cp\u003e\u0026nbsp;\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u003cstrong\u003e50.001 - 86.070\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e1\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e2\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e5\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e7\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e(53.20%) 8\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd rowspan=\"4\" width=\"76\"\u003e\n\u003cp\u003ePrincipal eigen value =\u0026nbsp;4.118,\u003c/p\u003e\n\u003cp\u003eConsistency Ratio CR\u0026nbsp;=\u0026nbsp;4.3%\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"57\"\u003e\n\u003cp\u003e7858\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"85\"\u003e\n\u003cp\u003e101\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"76\"\u003e\n\u003cp\u003e25.856\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u003cstrong\u003e30.001 - 50.00\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e0.5\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e1\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e3\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e3\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e(27.30%) 6\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"57\"\u003e\n\u003cp\u003e5793\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"85\"\u003e\n\u003cp\u003e62\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"76\"\u003e\n\u003cp\u003e19.062\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u003cstrong\u003e15.001 - 30.000\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e0.2\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e0.33\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e1\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e3\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e(12.80%) 5\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"57\"\u003e\n\u003cp\u003e12149\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"85\"\u003e\n\u003cp\u003e134\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"76\"\u003e\n\u003cp\u003e39.976\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u003cstrong\u003e4.520 - 15.000\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e0.14\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e0.33\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e0.33\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e1\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e(6.70%) 3\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"57\"\u003e\n\u003cp\u003e4591\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"85\"\u003e\n\u003cp\u003e66\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"76\"\u003e\n\u003cp\u003e15.106\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd rowspan=\"5\" width=\"38\"\u003e\n\u003cp\u003e\u003cstrong\u003eWL\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u0026nbsp;\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u003cstrong\u003e501 \u0026ndash; 1790\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e(very high)\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u003cstrong\u003e251 \u0026ndash; 500\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e(High)\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u003cstrong\u003e101 \u0026ndash; 250\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e(Moderate)\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u003cstrong\u003e0 \u0026ndash; 100\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e(Low)\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u0026nbsp;\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"76\"\u003e\n\u003cp\u003e\u0026nbsp;\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"57\"\u003e\n\u003cp\u003e\u0026nbsp;\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"85\"\u003e\n\u003cp\u003e\u0026nbsp;\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"76\"\u003e\n\u003cp\u003e\u0026nbsp;\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u003cstrong\u003e501 - 1790\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e1\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e3\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e4\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e7\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e(56.30%) 8\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd rowspan=\"4\" width=\"76\"\u003e\n\u003cp\u003ePrincipal eigen value =\u0026nbsp;4.063,\u003c/p\u003e\n\u003cp\u003eConsistency Ratio CR\u0026nbsp;=\u0026nbsp;2.3%\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"57\"\u003e\n\u003cp\u003e2616\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"85\"\u003e\n\u003cp\u003e30\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"76\"\u003e\n\u003cp\u003e8.608\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u003cstrong\u003e251 - 500\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e0.33\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e1\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e2\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e3\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e(22.30%) 6\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"57\"\u003e\n\u003cp\u003e11800\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"85\"\u003e\n\u003cp\u003e142\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"76\"\u003e\n\u003cp\u003e38.827\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u003cstrong\u003e101 - 250\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e0.25\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e0.5\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e1\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e3\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e(14.80%) 4\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"57\"\u003e\n\u003cp\u003e4737\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"85\"\u003e\n\u003cp\u003e56\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"76\"\u003e\n\u003cp\u003e15.587\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u003cstrong\u003e0 - 100\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e0.14\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e0.33\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e0.33\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e1\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e(6.70%)3\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"57\"\u003e\n\u003cp\u003e11238\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"85\"\u003e\n\u003cp\u003e134\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"76\"\u003e\n\u003cp\u003e36.978\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd rowspan=\"5\" width=\"38\"\u003e\n\u003cp\u003e\u003cstrong\u003eCVR\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u0026nbsp;\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u003cstrong\u003e86.401 - 87.563\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e(Low)\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u003cstrong\u003e87.564 - 92.888\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e(Moderate)\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u003cstrong\u003e92.889 - 101.239\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e(High)\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u003cstrong\u003e101.240 - 108.006\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e(Very high)\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u0026nbsp;\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"76\"\u003e\n\u003cp\u003e\u0026nbsp;\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"57\"\u003e\n\u003cp\u003e\u0026nbsp;\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"85\"\u003e\n\u003cp\u003e\u0026nbsp;\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"76\"\u003e\n\u003cp\u003e\u0026nbsp;\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u003cstrong\u003e86.401 - 87.563\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e1\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e2\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e4\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e7\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e(52.50%) 7\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd rowspan=\"4\" width=\"76\"\u003e\n\u003cp\u003ePrincipal eigen value =\u0026nbsp;4.050,\u003c/p\u003e\n\u003cp\u003eConsistency Ratio CR\u0026nbsp;=\u0026nbsp;1.8%\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"57\"\u003e\n\u003cp\u003e4293\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"85\"\u003e\n\u003cp\u003e87\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"76\"\u003e\n\u003cp\u003e14.126\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u003cstrong\u003e87.564 - 92.888\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e0.5\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e1\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e2\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e3\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e(25.40%) 6\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"57\"\u003e\n\u003cp\u003e6896\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"85\"\u003e\n\u003cp\u003e124\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"76\"\u003e\n\u003cp\u003e22.691\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u003cstrong\u003e92.889 - 101.239\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e0.25\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e0.5\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e1\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e3\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e(15.20%) 5\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"57\"\u003e\n\u003cp\u003e8816\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"85\"\u003e\n\u003cp\u003e78\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"76\"\u003e\n\u003cp\u003e29.009\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u003cstrong\u003e101.240 - 108.006\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e0.14\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e0.33\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e0.33\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e1\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e(6.90%) 3\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"57\"\u003e\n\u003cp\u003e10386\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"85\"\u003e\n\u003cp\u003e73\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"76\"\u003e\n\u003cp\u003e34.175\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd rowspan=\"5\" width=\"38\"\u003e\n\u003cp\u003e\u003cstrong\u003eVS\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u0026nbsp;\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u003cstrong\u003e7.1 \u0026ndash; 12\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e(Very High)\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u003cstrong\u003e4.1 \u0026ndash; 7\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e(High)\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u003cstrong\u003e1.1 - 4\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e(Moderate)\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u003cstrong\u003e0 \u0026ndash; 1\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e(Low)\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u0026nbsp;\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"76\"\u003e\n\u003cp\u003e\u0026nbsp;\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"57\"\u003e\n\u003cp\u003e\u0026nbsp;\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"85\"\u003e\n\u003cp\u003e\u0026nbsp;\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"76\"\u003e\n\u003cp\u003e\u0026nbsp;\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u003cstrong\u003e7.1 - 12\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e1\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e3\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e4\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e8\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e(57.20%) 8\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd rowspan=\"4\" width=\"76\"\u003e\n\u003cp\u003ePrincipal eigen value =\u0026nbsp;4.052,\u003c/p\u003e\n\u003cp\u003eConsistency Ratio CR\u0026nbsp;=\u0026nbsp;1.9%\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"57\"\u003e\n\u003cp\u003e5051\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"85\"\u003e\n\u003cp\u003e60\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"76\"\u003e\n\u003cp\u003e16.669\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u003cstrong\u003e4.1 - 7\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e0.33\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e1\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e2\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e3\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e(22.00%) 7\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"57\"\u003e\n\u003cp\u003e2932\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"85\"\u003e\n\u003cp\u003e37\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"76\"\u003e\n\u003cp\u003e9.676\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u003cstrong\u003e1.1 - 4\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e0.25\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e0.5\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e1\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e3\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e(14.50%) 5\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"57\"\u003e\n\u003cp\u003e8614\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"85\"\u003e\n\u003cp\u003e176\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"76\"\u003e\n\u003cp\u003e28.428\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u003cstrong\u003e0 - 1\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e0.12\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e0.33\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e0.33\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e1\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e(6.30%) 3\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"57\"\u003e\n\u003cp\u003e13704\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"85\"\u003e\n\u003cp\u003e92\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"76\"\u003e\n\u003cp\u003e45.226\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd rowspan=\"5\" width=\"38\"\u003e\n\u003cp\u003e\u003cstrong\u003eRF\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u0026nbsp;\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u003cstrong\u003e23-29\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e(very high)\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u003cstrong\u003e16-22\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e(High)\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u003cstrong\u003e8-15\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e(Moderate)\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u003cstrong\u003e0-7\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e(Low)\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u0026nbsp;\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"76\"\u003e\n\u003cp\u003e\u0026nbsp;\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"57\"\u003e\n\u003cp\u003e\u0026nbsp;\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"85\"\u003e\n\u003cp\u003e\u0026nbsp;\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"76\"\u003e\n\u003cp\u003e\u0026nbsp;\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u003cstrong\u003e23-29\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e1\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e3\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e4\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e8\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e(56.40%) 8\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd rowspan=\"4\" width=\"76\"\u003e\n\u003cp\u003ePrincipal eigen value =\u0026nbsp;4.048,\u003c/p\u003e\n\u003cp\u003eConsistency Ratio CR\u0026nbsp;=\u0026nbsp;1.7%\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"57\"\u003e\n\u003cp\u003e2788\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"85\"\u003e\n\u003cp\u003e31\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"76\"\u003e\n\u003cp\u003e9.174\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u003cstrong\u003e16-22\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e0.33\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e1\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e2\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e5\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e(24.20%) 6\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"57\"\u003e\n\u003cp\u003e2258\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"85\"\u003e\n\u003cp\u003e26\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"76\"\u003e\n\u003cp\u003e7.430\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u003cstrong\u003e8-15\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e0.25\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e0.5\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e1\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e3\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e(13.90%) 4\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"57\"\u003e\n\u003cp\u003e14813\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"85\"\u003e\n\u003cp\u003e177\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"76\"\u003e\n\u003cp\u003e48.741\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u003cstrong\u003e0-7\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e0.12\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e0.2\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e0.33\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e1\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e(5.40%) 3\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"57\"\u003e\n\u003cp\u003e10532\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"85\"\u003e\n\u003cp\u003e128\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"76\"\u003e\n\u003cp\u003e34.655\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd rowspan=\"5\" width=\"38\"\u003e\n\u003cp\u003e\u003cstrong\u003ePD\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u0026nbsp;\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u003cstrong\u003e44 \u0026ndash; 350\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e(Low)\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u003cstrong\u003e351 \u0026ndash; 700\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e(Moderate)\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u003cstrong\u003e701 \u0026ndash; 1000\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e(High)\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u003cstrong\u003e1001 \u0026ndash; 1313\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e(Very High)\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u0026nbsp;\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"76\"\u003e\n\u003cp\u003e\u0026nbsp;\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"57\"\u003e\n\u003cp\u003e\u0026nbsp;\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"85\"\u003e\n\u003cp\u003e\u0026nbsp;\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"76\"\u003e\n\u003cp\u003e\u0026nbsp;\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u003cstrong\u003e44 - 350\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e1\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e2\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e4\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e5\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e(50.70%) 8\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd rowspan=\"4\" width=\"76\"\u003e\n\u003cp\u003ePrincipal eigen value =\u0026nbsp;4.021,\u003c/p\u003e\n\u003cp\u003eConsistency Ratio CR\u0026nbsp;=\u0026nbsp;0.8%\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"57\"\u003e\n\u003cp\u003e10642\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"85\"\u003e\n\u003cp\u003e108\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"76\"\u003e\n\u003cp\u003e35.017\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u003cstrong\u003e351 - 700\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e0.5\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e1\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e2\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e3\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e(26.40%) 6\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"57\"\u003e\n\u003cp\u003e15875\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"85\"\u003e\n\u003cp\u003e207\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"76\"\u003e\n\u003cp\u003e51.940\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u003cstrong\u003e701 - 1000\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e0.25\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e0.5\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e1\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e2\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e(14.30%) 4\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"57\"\u003e\n\u003cp\u003e3651\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"85\"\u003e\n\u003cp\u003e43\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"76\"\u003e\n\u003cp\u003e12.013\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u003cstrong\u003e1001 - 1313\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e0.2\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e0.33\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e0.5\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e1\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e(8.60%) 3\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"57\"\u003e\n\u003cp\u003e313\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"85\"\u003e\n\u003cp\u003e04\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"76\"\u003e\n\u003cp\u003e1.030\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd rowspan=\"5\" width=\"38\"\u003e\n\u003cp\u003e\u003cstrong\u003ePGR\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u0026nbsp;\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u003cstrong\u003e20.001 - 24.440\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e(Very high)\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u003cstrong\u003e15.001 - 20.000\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e(High)\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u003cstrong\u003e10.001 - 15.000\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e(Moderate)\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u003cstrong\u003e5.210 - 10.000\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e(Low)\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u0026nbsp;\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"76\"\u003e\n\u003cp\u003e\u0026nbsp;\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"57\"\u003e\n\u003cp\u003e\u0026nbsp;\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"85\"\u003e\n\u003cp\u003e\u0026nbsp;\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"76\"\u003e\n\u003cp\u003e\u0026nbsp;\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u003cstrong\u003e20.001 - 24.440\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e1\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e2\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e4\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e5\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e(51.20%) 8\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd rowspan=\"4\" width=\"76\"\u003e\n\u003cp\u003e\u0026nbsp;\u003c/p\u003e\n\u003cp\u003ePrincipal eigen value =\u0026nbsp;4.047,\u003c/p\u003e\n\u003cp\u003eConsistency Ratio CR\u0026nbsp;=\u0026nbsp;1.7%\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"57\"\u003e\n\u003cp\u003e8213\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"85\"\u003e\n\u003cp\u003e98\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"76\"\u003e\n\u003cp\u003e27.024\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u003cstrong\u003e15.001 - 20.000\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e0.5\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e1\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e2\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e2\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e(24.40%) 6\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"57\"\u003e\n\u003cp\u003e11272\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"85\"\u003e\n\u003cp\u003e138\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"76\"\u003e\n\u003cp\u003e37.090\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u003cstrong\u003e10.001 - 15.000\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e0.25\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e0.5\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e1\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e2\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e(14.60%) 4\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"57\"\u003e\n\u003cp\u003e6737\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"85\"\u003e\n\u003cp\u003e75\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"76\"\u003e\n\u003cp\u003e22.168\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u003cstrong\u003e5.210 - 10.000\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e0.2\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e0.5\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e0.5\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e1\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e(9.80%) 2\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"57\"\u003e\n\u003cp\u003e4169\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"85\"\u003e\n\u003cp\u003e51\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"76\"\u003e\n\u003cp\u003e13.718\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd rowspan=\"5\" width=\"38\"\u003e\n\u003cp\u003e\u003cstrong\u003eLR\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u0026nbsp;\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u003cstrong\u003e79.31-88.71\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e(Very high)\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u003cstrong\u003e72.64-79.30\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e(High)\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u003cstrong\u003e65.38-72.63\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e(Moderate)\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u003cstrong\u003e58.34-65.37\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e(Low)\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u0026nbsp;\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"76\"\u003e\n\u003cp\u003e\u0026nbsp;\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"57\"\u003e\n\u003cp\u003e\u0026nbsp;\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"85\"\u003e\n\u003cp\u003e\u0026nbsp;\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"76\"\u003e\n\u003cp\u003e\u0026nbsp;\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u003cstrong\u003e79.31-88.71\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e1\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e2\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e3\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e5\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e(48.80%)8\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd rowspan=\"4\" width=\"76\"\u003e\n\u003cp\u003ePrincipal eigen value =\u0026nbsp;4.041,\u003c/p\u003e\n\u003cp\u003eConsistency Ratio CR\u0026nbsp;=\u0026nbsp;1.5%\u0026nbsp;\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"57\"\u003e\n\u003cp\u003e313\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"85\"\u003e\n\u003cp\u003e47\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"76\"\u003e\n\u003cp\u003e1.030\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u003cstrong\u003e72.64-79.30\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e0.5\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e1\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e2\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e2\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e(25.20%) 6\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"57\"\u003e\n\u003cp\u003e4704\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"85\"\u003e\n\u003cp\u003e116\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"76\"\u003e\n\u003cp\u003e15.478\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u003cstrong\u003e65.38-72.63\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e0.33\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e0.5\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e1\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e2\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e(16.10%) 4\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"57\"\u003e\n\u003cp\u003e8539\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"85\"\u003e\n\u003cp\u003e147\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"76\"\u003e\n\u003cp\u003e28.097\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u003cstrong\u003e58.34-65.37\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e0.2\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e0.5\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e0.5\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e1\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e(10.00%) 2\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"57\"\u003e\n\u003cp\u003e16835\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"85\"\u003e\n\u003cp\u003e50\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"76\"\u003e\n\u003cp\u003e55.395\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd rowspan=\"5\" width=\"38\"\u003e\n\u003cp\u003e\u003cstrong\u003eRRD\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u0026nbsp;\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u003cstrong\u003e300.01 - 482.69\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e(Very High)\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u003cstrong\u003e70.01 - 300.00\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e(High)\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u003cstrong\u003e30.01 - 70.00\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e(Moderate)\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u003cstrong\u003e0 - 30.00\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e(Low)\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u0026nbsp;\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"76\"\u003e\n\u003cp\u003e\u0026nbsp;\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"57\"\u003e\n\u003cp\u003e\u0026nbsp;\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"85\"\u003e\n\u003cp\u003e\u0026nbsp;\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"76\"\u003e\n\u003cp\u003e\u0026nbsp;\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u003cstrong\u003e300.01 - 482.69\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e1\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e2\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e3\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e9\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e(52.70%) 8\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd rowspan=\"4\" width=\"76\"\u003e\n\u003cp\u003ePrincipal eigen value =\u0026nbsp;4.021,\u003c/p\u003e\n\u003cp\u003eConsistency Ratio CR\u0026nbsp;=\u0026nbsp;0.8%\u0026nbsp;\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"57\"\u003e\n\u003cp\u003e156\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"85\"\u003e\n\u003cp\u003e01\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"76\"\u003e\n\u003cp\u003e0.513\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u003cstrong\u003e70.01 - 300.00\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e0.5\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e1\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e1\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e3\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e(21.50%) 6\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"57\"\u003e\n\u003cp\u003e1337\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"85\"\u003e\n\u003cp\u003e16\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"76\"\u003e\n\u003cp\u003e4.399\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u003cstrong\u003e30.01 - 70.00\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e0.33\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e1\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e1\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e3\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e(19.30%) 5\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"57\"\u003e\n\u003cp\u003e9701\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"85\"\u003e\n\u003cp\u003e232\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"76\"\u003e\n\u003cp\u003e31.921\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e\u003cstrong\u003e0 - 30.00\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e0.11\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e0.33\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e0.33\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e1\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"66\"\u003e\n\u003cp\u003e(6.40%) 3\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"57\"\u003e\n\u003cp\u003e19197\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"85\"\u003e\n\u003cp\u003e113\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd width=\"76\"\u003e\n\u003cp\u003e63.167\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003c/tbody\u003e\n\u003c/table\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec17\" class=\"Section2\"\u003e\n\u003ch2\u003e4.3 Analysis of different pre-requisites of machine learning models\u003c/h2\u003e\n\u003cp\u003eBefore initialization of machine learning models, we develop 10-fold repeated cross validation framework. Our objective with the cross validation is to optimize the size of the tree. After plotting the 1-P value threshold versus accuracy (repeated cross validation) we observe that how accuracy is maximized into a relatively less complex tree (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e7\u003c/span\u003e). The final value used for the model was mincriterion 0.01; which was used to best tune the model. In the final model total 16 terminal nodes were created in the regression trees (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e8\u003c/span\u003e). Rules were illustrated for this ctree model in Appendix. 1. Within 16 terminal nodes, the model predicted criterion 1 (Very High to High TPZ) 9 times. Final Bagged CART (treebag model) model was created by using 25 bootstrap replications. For the random forest model (rf) the final accuracy was used to select the optimal model using the largest value. The final value for the rf model is mtry\u0026thinsp;=\u0026thinsp;11 (best tuned). For this final model rf model, number of trees are 500 and the number of variables tried to each split is 6. The out of the bag error (OOB) estimate for the model is 17.25%. For the cforest model, the final accuracy was used to select the optimal model using the largest value mtry\u0026thinsp;=\u0026thinsp;11 (best tuned) with the 500 number of trees in the final model. The final tuned gradient boosted model was created with Bernoulli Loss function with 50 iterations. It has a 50 number of trees, interaction depth 2, shrinkage 0.1 and minobsinnode 10.\u003c/p\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec18\" class=\"Section2\"\u003e\n\u003ch2\u003e4.4 Variable importance analysis\u003c/h2\u003e\n\u003cp\u003eTourism Potentiality is a multifaceted conception, which are dependent on several criteria (Raha and Gayen \u003cspan class=\"CitationRef\"\u003e2021\u003c/span\u003e). Therefore, it is really essential to determine tourism potentiality effective factors and their contribution. After train each model, Variable Importance Plot (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e9\u003c/span\u003e) for each model was assessed. For the ctree model (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e9\u003c/span\u003ea), the FA bears the highest importance. The FA was followed by LR, WL, RF, RL, CVR, RRD, AS, VS and PGR (least importance). For the treebag model (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e9\u003c/span\u003eb), RRD was marked with the highest importance and the PD was identified with the least importance. RRD was followed by LR, FA, RL, AS, CVR, VS, RF, PGR, WL and PD. In the rf model (as shown in Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e9\u003c/span\u003ec), the feature RRD was identified as having the greatest importance, with FA, LR, RL, AS, VS, PD, RF, CVR, WL, and PGR following in descending order of importance. For the cforest model (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e9\u003c/span\u003ed), FA was marked with the highest importance followed by RRD, LR, PD, RF, RL, PGR, CVR, VS, WL and AS. The FA was also found with the highest importance for the gbm model (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e9\u003c/span\u003ee). Here WL was found with the least importance. RRD and LR were also found with substantial importance for the gbm model. For the TPZ Ensemble Model (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e9\u003c/span\u003ef), rf model was found the highest importance followed by gbm, treebag, ctree and cforest model.\u003c/p\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec19\" class=\"Section2\"\u003e\n\u003ch2\u003e4.5 Tourism Potentiality assessment models\u003c/h2\u003e\n\u003cp\u003eThe tourism potentiality assessment models by applying five machine learning (i.e., ctree, treebag, rf, cforest, and gbm) and one ensemble model were presented in this research (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e10\u003c/span\u003ea,\u003cspan class=\"InternalRef\"\u003e10\u003c/span\u003eb,\u003cspan class=\"InternalRef\"\u003e11\u003c/span\u003ea,\u003cspan class=\"InternalRef\"\u003e11\u003c/span\u003eb, \u003cspan class=\"InternalRef\"\u003e12\u003c/span\u003ea, \u003cspan class=\"InternalRef\"\u003e12\u003c/span\u003eb). The all models show very high, high and moderate to low tourism potentiality followed by natural break strategy (Gayen et al. \u003cspan class=\"CitationRef\"\u003e2019\u003c/span\u003e). Higher value indicates higher tourism potentiality.\u003c/p\u003e\n\u003cp\u003eUsing Ensemble model approximately, 35.42% area was marked as the high to very high tourism potentiality. On he other hand, approximately 64.58% area was identified as the moderate to low tourism potentiality. Using the Bagged CART model (treebag), approximately 53.61% area was identified as the high to very high tourism potentiality and 46.39% area was demarcated as moderate to low tourism potentiality. Using the Conditional Inference Tree (ctree) model, approximately 50.62% area was demarcated as high to very high tourism potentiality and remaining 49.38% area was marked as the moderate to low tourism potentiality. Using Conditional Inference Random Forest (cforest) model, near about 40.52% and 59.48% area were identified as the high to very high and moderate to low tourism potentiality respectively. Using the rf model, approximately 28.24% area was identified as very high to high tourism potentiality. Remaining 71.76% area was identified with moderate to low tourism potentialities. Approximately 36.5% area was demarcated as High to Very High tourism potentialities by the gradient boosting (gbm model) model. On the contrary, 63.5% area was marked with moderate to low tourism potentiality (Table\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e4\u003c/span\u003e).\u003c/p\u003e\n\u003cdiv class=\"gridtable\"\u003e\n\u003ctable id=\"Tab4\" border=\"1\"\u003e\u003ccaption\u003e\n\u003cdiv class=\"CaptionNumber\"\u003eTable 4\u003c/div\u003e\n\u003cdiv class=\"CaptionContent\"\u003e\n\u003cp\u003ePixel count with percentage area by different models\u003c/p\u003e\n\u003c/div\u003e\n\u003c/caption\u003e\n\u003cthead\u003e\n\u003ctr\u003e\n\u003cth align=\"left\"\u003e\u0026nbsp;\u003c/th\u003e\n\u003cth colspan=\"4\" align=\"left\"\u003e\n\u003cp\u003ePixel Count\u003c/p\u003e\n\u003c/th\u003e\n\u003cth colspan=\"4\" align=\"left\"\u003e\n\u003cp\u003e% Area\u003c/p\u003e\n\u003c/th\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003cth align=\"left\"\u003e\n\u003cp\u003eModels\u003c/p\u003e\n\u003c/th\u003e\n\u003cth align=\"left\"\u003e\n\u003cp\u003eModerate to Low\u003c/p\u003e\n\u003c/th\u003e\n\u003cth align=\"left\"\u003e\n\u003cp\u003eHigh\u003c/p\u003e\n\u003c/th\u003e\n\u003cth align=\"left\"\u003e\n\u003cp\u003eVery High\u003c/p\u003e\n\u003c/th\u003e\n\u003cth align=\"left\"\u003e\n\u003cp\u003eTotal\u003c/p\u003e\n\u003c/th\u003e\n\u003cth align=\"left\"\u003e\n\u003cp\u003eModerate to Low\u003c/p\u003e\n\u003c/th\u003e\n\u003cth align=\"left\"\u003e\n\u003cp\u003eHigh\u003c/p\u003e\n\u003c/th\u003e\n\u003cth align=\"left\"\u003e\n\u003cp\u003eVery High\u003c/p\u003e\n\u003c/th\u003e\n\u003cth align=\"left\"\u003e\n\u003cp\u003eTotal\u003c/p\u003e\n\u003c/th\u003e\n\u003c/tr\u003e\n\u003c/thead\u003e\n\u003ctbody\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e\u003cstrong\u003eEnsemble\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e19339\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e8353\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e2253\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e29945\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e64.58\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e27.89\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e7.53\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e100\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e\u003cstrong\u003eBagged CART\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e13891\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e13152\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e2902\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e29945\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e46.39\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e43.92\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e9.69\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e100\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e\u003cstrong\u003eConditional Inference Random Forest\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e17811\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e11608\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e526\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e29945\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e59.48\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e38.76\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e1.76\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e100\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e\u003cstrong\u003eConditional Inference Tree\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e14787\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e11716\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e3442\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e29945\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e49.38\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e39.13\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e11.49\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e100\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e\u003cstrong\u003eRandom Forest\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e21489\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e7350\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e1106\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e29945\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e71.76\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e24.55\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e3.69\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e100\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e\u003cstrong\u003eGradient Boosting\u003c/strong\u003e\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e19015\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e10488\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e442\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e29945\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e63.5\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e35.02\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e1.48\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e100\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003c/tbody\u003e\n\u003c/table\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec20\" class=\"Section3\"\u003e\n\u003ch2\u003e4.4.1 Very High (VH) and High (H) TPZ\u003c/h2\u003e\n\u003cp\u003eThe detailed characteristics of this zone was briefly discussed as follows:\u003c/p\u003e\n\u003cul\u003e\n\u003cli\u003e\n\u003cp\u003eVH and H tourism potentialities dominated in the Middle portions, south-eastern and southern sections of the Assam. Over 50% area of Golpara, Hailakandi, Jorhat, Kamrup, Kamrup Metropolitan, Karbi Anglong, Karimganj, Lakhimpur, Naogaon, Sivasagar, Tinsukia and West Karbi Anglong districts were marked with Very high potentials for tourism (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e10\u003c/span\u003ea,\u003cspan class=\"InternalRef\"\u003e10\u003c/span\u003eb,\u003cspan class=\"InternalRef\"\u003e11\u003c/span\u003ea,\u003cspan class=\"InternalRef\"\u003e11\u003c/span\u003eb, \u003cspan class=\"InternalRef\"\u003e12\u003c/span\u003ea, \u003cspan class=\"InternalRef\"\u003e12\u003c/span\u003eb).\u003c/p\u003e\n\u003c/li\u003e\n\u003cli\u003e\n\u003cp\u003eThese sections were marked with the moderate to high relief structure (80.01 meter to 1971 meter) (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e4\u003c/span\u003ea), which creates a wider viewshed of the location and this attracts tourists conveniently and easily (Huff and Tingley \u003cspan class=\"CitationRef\"\u003e2015\u003c/span\u003e).\u003c/p\u003e\n\u003c/li\u003e\n\u003cli\u003e\n\u003cp\u003eHigher number of reserved forests (8 to 29 numbers, Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e4\u003c/span\u003ed) and larger forest area (30.001\u0026ndash;86.07% area, Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e4\u003c/span\u003ec) in these sections created amazing ambience for the tourism activities.\u003c/p\u003e\n\u003c/li\u003e\n\u003cli\u003e\n\u003cp\u003eIt is the region, which is dominated by the wetlands (i.e., 251 to 1790 Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e5\u003c/span\u003ea) and rainfall variation (i.e., 92.889\u0026ndash;108.006%, Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e5\u003c/span\u003eb). A wetland is an area where the ground is continuously or periodically flooded by water, whether it be salty, pure, or a combination of both. The seasonal variation of rainfall accelerates the vegetation pattern and forest cover of a particular region. The seasonal variation creates the rthymic diversity of forest cover (Ghazoul \u003cspan class=\"CitationRef\"\u003e2016\u003c/span\u003e).\u003c/p\u003e\n\u003c/li\u003e\n\u003cli\u003e\n\u003cp\u003eThis section has a higher road and rail density (greater than 70, Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e6\u003c/span\u003ed), which accelerates the connectivity of the region. Tourists can reach the tourist destination more relatedly and easily by accessible roads and rails (Holloway and Humphreys \u003cspan class=\"CitationRef\"\u003e2022\u003c/span\u003e).\u003c/p\u003e\n\u003c/li\u003e\n\u003cli\u003e\n\u003cp\u003eThe population density (351 to 1000 persons/ sq.km., Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e6\u003c/span\u003eb), population growth rate (10\u0026ndash;24%, Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e9\u003c/span\u003ea) and the literacy rate (72.64\u0026ndash;88.71%, Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e9\u003c/span\u003ec) are higher in these sections of the study area. The literacy rate creates awareness about a balanced population structure; therefore, it creates favorable environment for the tourism activities (Getz and Page \u003cspan class=\"CitationRef\"\u003e2019\u003c/span\u003e).\u003c/p\u003e\n\u003c/li\u003e\n\u003c/ul\u003e\n\u003cp\u003eIrrespective of the above issues, these portions are affected by comparatively low pollution levels, as it is dominated by higher number of reserved forests and forest area (Singh et al. \u003cspan class=\"CitationRef\"\u003e2020\u003c/span\u003e). As a result, these portions are marked with higher potentials for the tourism.\u003c/p\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec21\" class=\"Section3\"\u003e\n\u003ch2\u003e4.4.2 Moderate (M) to Low (L) TPZ\u003c/h2\u003e\n\u003cp\u003eThe detailed characteristics of this zone was briefly discussed as follows:\u003c/p\u003e\n\u003cul\u003e\n\u003cli\u003e\n\u003cp\u003eThe upper Assam and the upper north-eastern portions were marked with the low tourism potentials. Over 50% area of Baksa, Barpeta, Bongaigaon, Chirang, Darrang, Dhemaji, Dhubri, Dibrugarh, Golaghat, Kokrajhar, Morigaon, Nalbari, Sontipur, and Udayguri were identified under the moderate to low tourism potentials (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e10\u003c/span\u003ea,\u003cspan class=\"InternalRef\"\u003e10\u003c/span\u003eb,\u003cspan class=\"InternalRef\"\u003e11\u003c/span\u003ea,\u003cspan class=\"InternalRef\"\u003e11\u003c/span\u003eb, \u003cspan class=\"InternalRef\"\u003e12\u003c/span\u003ea, \u003cspan class=\"InternalRef\"\u003e12\u003c/span\u003eb).\u003c/p\u003e\n\u003c/li\u003e\n\u003cli\u003e\n\u003cp\u003eComparatively flat terrain (1 to 300meter, Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e4\u003c/span\u003ea,\u003cspan class=\"InternalRef\"\u003e4\u003c/span\u003eb,\u003cspan class=\"InternalRef\"\u003e4\u003c/span\u003ec) was marked in these sections which does not create any picturesque beauty and do create a low viewshed of a particular region.\u003c/p\u003e\n\u003c/li\u003e\n\u003cli\u003e\n\u003cp\u003eNear about 5\u0026ndash;30% forest area (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e4\u003c/span\u003ed) and comparatively low number of reserved forests (0 to 15 numbers). These factors lower the ambience, vibrations and motivations of the tourist activities.\u003c/p\u003e\n\u003c/li\u003e\n\u003cli\u003e\n\u003cp\u003eThese sections are scarced of wetlands (i.e., 0 to 250 numbers Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e5\u003c/span\u003ea) and these sections are marked with a very low amount of rainfall variation (i.e., 86.401\u0026ndash;92.888%, Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e5\u003c/span\u003eb).\u003c/p\u003e\n\u003c/li\u003e\n\u003cli\u003e\n\u003cp\u003eModerate to high population density (i.e., 351 to 1000 persons/square km., Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e6\u003c/span\u003eb) and growth rate (i.e., 10.001\u0026ndash;20.000%) are also marked in these portions of the study area (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e6\u003c/span\u003ea). Literacy rates (i.e., 58.34\u0026ndash;72.63%, Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e6\u003c/span\u003ec) are also comparatively poor.\u003c/p\u003e\n\u003c/li\u003e\n\u003cli\u003e\n\u003cp\u003eThese portions have low to moderate RRD values (0 to 70) (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e6\u003c/span\u003ed).\u003c/p\u003e\n\u003c/li\u003e\n\u003c/ul\u003e\n\u003cp\u003eApart from the above specified issues, these portions are affected by comparatively high pollution levels, as it is crowded with different types of industries. These sections are over congested. Therefore, these sections are not attracted by tourists hence having a comparatively low tourism potential.\u003c/p\u003e\n\u003c/div\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec22\" class=\"Section2\"\u003e\n\u003ch2\u003e4.3 Validation\u003c/h2\u003e\n\u003cp\u003eThe Ensemble model was appeared with the highest Kappa (0.81), accuracy (0.93 value) and AUC-ROC (96.9% area) values. Based on AUC-ROC measurement, the Ensemble model was followed by cforest model (89.7% AUC), rf model (79.9% AUC), gbm model (78.8% AUC values), ctree model (74.9% AUC values) and treebag model (74.01% AUC-ROC). According to the quality criteria of AUC, the performance of ensemble model was appeared as \u0026lsquo;excellent\u0026rsquo;. The performance of cforest model was \u0026lsquo;accepted\u0026rsquo;. The performance of remaining models (i.e., ctree, treebag, gbm, and rf models) are \u0026lsquo;good\u0026rsquo; or \u0026lsquo;satisfactory\u0026rsquo;. Based on the accuracy and Kappa measurement, also Ensemble model outperformed other models. The Ensemble model was followed by rf model, gbm model, cforest model, ctree model and treebag model. For both cases, ctree and treebag are the worst performer. On the contrary, rf, cforest and gbm models performed well. ROC-AUC and accuracy plots were outlined in Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e13\u003c/span\u003e and Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e14\u003c/span\u003e and Table\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e5\u003c/span\u003e.\u003c/p\u003e\n\u003cdiv class=\"gridtable\"\u003e\n\u003ctable id=\"Tab5\" border=\"1\"\u003e\u003ccaption\u003e\n\u003cdiv class=\"CaptionNumber\"\u003eTable 5\u003c/div\u003e\n\u003cdiv class=\"CaptionContent\"\u003e\n\u003cp\u003eAccuracy of different models (predictive accuracy)\u003c/p\u003e\n\u003c/div\u003e\n\u003c/caption\u003e\n\u003cthead\u003e\n\u003ctr\u003e\n\u003cth align=\"left\"\u003e\n\u003cp\u003eModel names\u003c/p\u003e\n\u003c/th\u003e\n\u003cth align=\"left\"\u003e\n\u003cp\u003eAccuracy\u003c/p\u003e\n\u003c/th\u003e\n\u003cth align=\"left\"\u003e\n\u003cp\u003eKappa\u003c/p\u003e\n\u003c/th\u003e\n\u003cth align=\"left\"\u003e\n\u003cp\u003eRank\u003c/p\u003e\n\u003cp\u003e(Based on Accuracy and Kappa)\u003c/p\u003e\n\u003c/th\u003e\n\u003cth align=\"left\"\u003e\n\u003cp\u003eModel Names\u003c/p\u003e\n\u003c/th\u003e\n\u003cth align=\"left\"\u003e\n\u003cp\u003eAUC_ROC Values\u003c/p\u003e\n\u003c/th\u003e\n\u003cth colspan=\"2\" align=\"left\"\u003e\n\u003cp\u003e95% Confidence Interval (For AUC)\u003c/p\u003e\n\u003c/th\u003e\n\u003cth align=\"left\"\u003e\n\u003cp\u003eRank (Based on AUC)\u003c/p\u003e\n\u003c/th\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003cth colspan=\"6\" align=\"left\"\u003e\u0026nbsp;\u003c/th\u003e\n\u003cth align=\"left\"\u003e\n\u003cp\u003eUpper\u003c/p\u003e\n\u003cp\u003eBound\u003c/p\u003e\n\u003c/th\u003e\n\u003cth align=\"left\"\u003e\n\u003cp\u003eLower\u003c/p\u003e\n\u003cp\u003eBound\u003c/p\u003e\n\u003c/th\u003e\n\u003cth align=\"left\"\u003e\u0026nbsp;\u003c/th\u003e\n\u003c/tr\u003e\n\u003c/thead\u003e\n\u003ctbody\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003eEnsemble\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e0.92\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e0.81\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e1\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003eEnsemble\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e96.9%\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e0.850\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e0.962\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e1\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003eConditional Inference Tree (cforest)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e0.81\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e0.62\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e4\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003eConditional Inference Tree (cforest)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e89.7%\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e0.723\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e0.878\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e2\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003eRandom Forest (rf)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e0.85\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e0.68\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e2\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003eRandom Forest (rf)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e79.9%\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e0.713\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e0.870\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e3\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003eStochastic Gradient Boosting (gbm)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e0.84\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e0.67\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e3\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003eStochastic Gradient Boosting (gbm)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e78.8%\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e0.693\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e0.854\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e4\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003eConditional Inference Tree (ctree)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e0.80\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e0.60\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e5\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003eConditional Inference Tree (ctree)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e74.9%\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e0.806\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e0.896\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e5\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003eBagged CART (treebag)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e0.79\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e0.57\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e6\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003eBagged CART (treebag)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e74.01%\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e0.693\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e0.895\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\" char=\".\"\u003e\n\u003cp\u003e6\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003c/tbody\u003e\n\u003c/table\u003e\n\u003c/div\u003e\n\u003c/div\u003e"},{"header":"5. Discussion","content":"\u003cp\u003eThe Ensemble method is the most accurate approach used in this research, according to the results of accuracy assessments. Combining the forecasts of various boot strapping base models such as decision trees, random forests, gradient boosting, ensemble methods eliminated the individual model's shortcomings and benefit from their advantages. Here the ensemble method helps to improve predictive performance by drawing on the combined wisdom of multiple models. Therefore, the formulation of ‘TPZ ensemble model’ enlarges a new direction in tourism research. In a move to be increasingly precise and effective, meticulous work was done on a new ensemble model of wide variety machine learning methodologies. This product, the TPZ ensemble model, has outstripped not only individual models but it is pure predictive ability in both the unparalleled accuracy metrics. Rigorous scrutiny showed that the TPZ ensemble model surpassed its constituent parts significantly, with an exceptional Area Under the Curve (AUC) value for example 96.9%. Also, its strong performance is evident in a Kappa coefficient of 0.81 - substantial agreement beyond pure chance. The most amazing thing is that the ensemble model, with its generalization capability still intact, actually gets 92% right. Further, the Ensemble methods helps in the decision-making process and helps choosing the best model from a variety and bunch of several ML models.\u003c/p\u003e \u003cp\u003eFurther, the rf and cforest models handled big datasets without variable deletion may have contributed to their strong performance in this study. According to Catani et al. (\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e2013\u003c/span\u003e), the rf model can handle nonlinearities between dominant factors and therefore provide good performance in the current context. With respectable performances, this model has also shown to be helpful in other research domains, including mapping groundwater potential, predicting wildfires, modeling sediment yield (Masselink et al., \u003cspan citationid=\"CR59\" class=\"CitationRef\"\u003e2017\u003c/span\u003e), and mapping landslide susceptibility (LSM) (Taalab et al. \u003cspan citationid=\"CR96\" class=\"CitationRef\"\u003e2018\u003c/span\u003e; Zhang et al. \u003cspan citationid=\"CR110\" class=\"CitationRef\"\u003e2017\u003c/span\u003e). The Gradient boosting Models also performed well in prediction accuracy may be attributed to the fact that it combines the predictions of several base estimators, typically decision trees. From the current study, the accuracy of the combined models was much higher compared to that of any individual model and also because decision trees are not sensitive to outliers. The high performance is scaled further by their ability to support large datasets and therefore can be used in applications with huge datasets. However, decision trees (like the Conditional Inference Tree and the Bagged CART models) have the ability to isolate outliers in different leaves, which keeps them from having a substantial impact on the performance of the model as a whole. Also, large data cannot be efficiently handled by ctree and Bagged CART model. They are also sensitive to outliers. Although in this research, their performance is quite satisfactory but in comparison to other models, the prediction accuracy, Kappa and AUC-ROC is quite low.\u003c/p\u003e "},{"header":"Conclusion","content":"\u003cp\u003eIn this research, an ensemble model and a variety of machine learning algorithms (i.e., Conditional Inference Tree, Bagged CART, Random Forest, Random Forest with Conditional Inference Tree, Gradient Boosting and one ensemble model) were used to predict the Tourism Potential Zones (TPZ) for the state of Assam. Here, first of all, we created a comprehensive Tourism Inventory Database from field research and Google Earth imagery, which produced 365 tourism points. Our study produced encouraging results by utilizing a wide range of tourism conditioning factors as independent variables, such as Relief, Aspect, Viewshed, Forest Area, Wetland, and socio-economic criteria like Population Density, Literacy Rate, and Road-Railway Density. The results clearly showed the good quality of the maps generated, with the best agreement between the ML models and tourism inventory data points. All of the spatially assessed TPZ maps show that Very High to high tourism potentialities dominated in the south-eastern and southern sections of the Assam. High to very high tourism potentials in this research are associated with moderate to low relief, higher viewshed, forest area, wetland, rainfall variation, higher literacy rate and better communication network (here Road and Railway density). Over 50% area of Golpara, Hailakandi, Jorhat, Kamrup, Kamrup Metropolitan, Karbi Anglong, Karimganj, Lakhimpur, Naogaon, Sivasagar, Tinsukia and West Karbi Anglong districts were marked with very high to high tourism potentials and Over 50% area of Baksa, Barpeta, Bongaigaon, Chirang, Darrang, Dhemaji, Dhubri, Dibrugarh, Golaghat, Kokrajhar, Morigaon, Nalbari, Sontipur, and Udayguri were identified under the lower tourism potentials. We found that all models performed admirably in terms of prediction accuracy after conducting a thorough analysis using Kappa, Accuracy and AUC-ROC method. TPZ ensemble model was proposed by combining other base models, and it is interesting to note that the ensemble model emerged as the frontrunner, showcasing superior metrics including the highest AUC (97.6%), Kappa (0.82), and accuracy (0.93) values.\u003c/p\u003e\u003cp\u003eThis research underscores the potential of machine learning and ensemble methods in predicting TPZs, furnishing valuable insights for decision-makers tasked with spearheading the development of tourism in Assam. By offering robust and nuanced predictions, our findings contribute to informed decision-making processes aimed at harnessing the rich tourism potential of the region, thereby fostering sustainable growth and prosperity.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eAcknowledgement\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAuthors expresses their gratitude to all faculty members of Department of Geography, Bhairab Ganguly College, who wholeheartedly support this research.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConflict of Interest\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThere is no conflict of interest regarding publication of this article.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding Source\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompliance with Ethical Standards\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe ethical approval is not applicable for this research.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eData availability statement\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eData are available from authors upon a reasonable request.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eApostolopoulos, D, Nikolakopoulos, K (2021) A review and meta-analysis of remote sensing data, GIS methods, materials and indices used for monitoring the coastline evolution over the last twenty years. Eur J Remote Sens 54:240-265. https://doi.org/10.1080/22797254.2021.1904293\u003c/li\u003e\n\u003cli\u003eBanik, S, Mukhopadhyay, M (2020) Model-based strategic planning for the development of community-based tourism: a case study of Ayodhya Hills in West Bengal, India. GeoJournal:1-17. https://doi.org/10.1007/s10708-020-10314-0\u003c/li\u003e\n\u003cli\u003eBast, H, Delling, D, Goldberg, A, M\u0026uuml;ller-Hannemann, M, Pajor, T, Sanders, P, \u0026hellip; Werneck R F (2016) Route Planning in Transportation Networks. In L. Kliemann \u0026amp; P. Sanders (Eds.), Algorithm Engineering: Selected Results and Surveys (pp. 19\u0026ndash;80). Cham: Springer International Publishing. https://doi.org/10.1007/978-3-319-49487-6_2\u003c/li\u003e\n\u003cli\u003eBi JW, Han TY, Li H (2022) International tourism demand forecasting with machine learning models: The power of the number of lagged inputs. Tourism Econ 28:621-645. https://doi.org/10.1177/1354816620976954\u003c/li\u003e\n\u003cli\u003eBlapp M, Mitas, O (2018) Creative tourism in Balinese rural communities. Current Issues in Tourism 21:1285\u0026ndash;1311. https://doi.org/10.1080/13683500.2017.1358701\u003c/li\u003e\n\u003cli\u003eBlockeel H (2011) Hypothesis Space Encyclopedia of Machine Learning, 511\u0026ndash;513.\u003c/li\u003e\n\u003cli\u003eBordoloi AK, Agarwal BK (2015) Tourism Potentiality in Tinsukia district of Upper Assam: An Analysis. International Journal of Management and Development Studies 4:375\u0026ndash;383. https://doi.org/10.53983/ijmds.v4i4.164\u003c/li\u003e\n\u003cli\u003eBreiman L (2001) Random forests. Machine learning 45:5-32.\u003c/li\u003e\n\u003cli\u003eCankurt S, Subasi A (2015) Developing tourism demand forecasting models using machine learning techniques with trend, seasonal, and cyclic components. Balkan Journal of Electrical and Computer Engineering 3:42-49. https://dergipark.org.tr/en/download/article-file/39709\u003c/li\u003e\n\u003cli\u003eCatani F, Lagomarsino D, Segoni S, Tofani V (2013) Landslide susceptibility estimation by random forests technique: sensitivity and scaling issues. Nat Hazard Earth Syst Sciences 13:2815-2831.\u003c/li\u003e\n\u003cli\u003eChaturvedi V, de Vries WT. (2021) Machine learning algorithms for urban land use planning: A review. Urban Science 5:68. https://doi.org/10.3390/urbansci5030068\u003c/li\u003e\n\u003cli\u003eChien C F, Ku C C, Lu Y Y (2023) Ensemble learning for demand forecast of After-Market spare parts to empower data-driven value chain and an empirical study. Computers \u0026amp; Industrial Engineering 185:109670. https://doi.org/10.1016/j.cie.2023.109670\u003c/li\u003e\n\u003cli\u003eChoden K, Wangchuk D, Smart! C (2018) Bhutan\u0026mdash;Culture Smart! The Essential Guide to Customs \u0026amp; Culture. Kuperard.\u003c/li\u003e\n\u003cli\u003eChoi S, Hur J (2020) An ensemble learner-based bagging model using past output data for photovoltaic forecasting. Energies 13:1438.\u003c/li\u003e\n\u003cli\u003eChowdary V M, Chakraborthy, D, Jeyaram, A, Murthy Y K, Sharma J R, Dadhwal VK (2013) Multi-criteria decision-making approach for watershed prioritization using analytic hierarchy process technique and GIS. Water Rsour Manage 27: 3555-3571. https://doi.org/10.1007/s11269-013-0364-6\u003c/li\u003e\n\u003cli\u003eClaveria O, Monte E, Torra S (2016) Combination forecasts of tourism demand with machine learning models. Appl Econ Letter 23:428-431. https://doi.org/10.1080/13504851.2015.1078441\u003c/li\u003e\n\u003cli\u003eCodrea PM, Bilașco Ș, Roșca S, Irimuș IA, Iuliu V, Rusu R, ... Sestras P (2022) The Integrated Assessment of Degraded Tourist Geomorphosites to Develop Sustainable Tourism: A Case Study of Grădina Zmeilor Geomorphosite, North-West Region, Romania. Appl Sci-Basel 12: 9816. https://doi.org/10.3390/app12199816\u003c/li\u003e\n\u003cli\u003eCouronn\u0026eacute; R, Probst P, Boulesteix AL (2018) Random forest versus logistic regression: a large-scale benchmark experiment. BMC Bioinformatics 19:1-14. https://doi.org/10.1186/s12859-018-2264-5\u003c/li\u003e\n\u003cli\u003eDanish M S S (2023) AI and expert insights for sustainable energy future. Energies 16: 3309. https://doi.org/10.3390/en16083309\u003c/li\u003e\n\u003cli\u003eDas S (2020) Flood susceptibility mapping of the Western Ghat coastal belt using multi-source geospatial data and analytical hierarchy process (AHP). Remote Sensing Applications: Society and Environment 20:100379. https://doi.org/10.1016/j.rsase.2020.100379\u003c/li\u003e\n\u003cli\u003eDeribew KT, Mihretu Y, Abreha G, Gemeda DO (2022) Spatial analysis of potential ecological sites in the northeastern parts of Ethiopia using multi-criteria decision-making models. Asia-Pacific Journal of Regional Science 6: 961\u0026ndash;991. https://doi.org/10.1007/s41685-022-00248-5\u003c/li\u003e\n\u003cli\u003eDong X, Yu Z, Cao W, Shi Y, Ma Q (2020) A survey on ensemble learning. Front Comput Sci-Chi 14:241-258.\u003c/li\u003e\n\u003cli\u003eFakfare P, Talawanich S, Wattanacharoensil W (2020) A scale development and validation on domestic tourists\u0026rsquo; motivation: The case of second-tier tourism destinations. Asia Pacific Journal of Tourism Research, 25:489\u0026ndash;504. https://doi.org/10.1080/10941665.2020.1745855\u003c/li\u003e\n\u003cli\u003eFox EW, Hill RA., Leibowitz SG., Olsen AR., Thornbrugh DJ., Weber MH. (2017) Assessing the accuracy and stability of variable selection methods for random forest modeling in ecology. Environ Monit Assess 189:1-20. https://doi.org/10.1007/s10661-017-6025-0\u003c/li\u003e\n\u003cli\u003eFu W (2017) Nonparametric Methods in Statistical Learning: Unbiasedness in Regression Trees, Survival Trees for Nonstandard Data and Estimating the Number of Clusters (Doctoral dissertation, New York University, Graduate School of Business Administration).\u003c/li\u003e\n\u003cli\u003eGanaie MA, Hu M, Malik AK., Tanveer M, Suganthan PN (2022) Ensemble deep learning: A review. Eng Appl of Artif Intel 115:105151. https://doi.org/10.1016/j.engappai.2022.105151\u003c/li\u003e\n\u003cli\u003eGayen A, Pourghasemi HR, Saha S, Keesstra S, Bai S (2019) Gully erosion susceptibility assessment and management of hazard-prone areas in India using different machine learning algorithms. Sci Total Environ, 668:124-138.\u003c/li\u003e\n\u003cli\u003eGetz D, Page SJ (2019) Theory, Research and Policy for Planned Events (4th ed.). London: Routledge. https://doi.org/10.4324/9780429023002\u003c/li\u003e\n\u003cli\u003eGevrey M, Dimopoulos I, Lek S (2003) Review and comparison of methods to study the contribution of variables in artificial neural network models. Ecol Model 160:249-264.\u003c/li\u003e\n\u003cli\u003eGhazoul J (2016) Dipterocarp Biology, Ecology, and Conservation. Oxford University Press.\u003c/li\u003e\n\u003cli\u003eGarg S, Sinha S, Kar AK, Mani M (2022) A review of machine learning applications in human resource management. International Journal of Productivity and Performance Management 71:1590-1610. https://doi.org/10.1108/IJPPM-08-2020-0427\u003c/li\u003e\n\u003cli\u003eGiorgi, F, Lionello P (2008) Climate change projections for the Mediterranean region. Global Planet Change 63:90\u0026ndash;104. https://doi.org/10.1016/j.gloplacha.2007.09.005\u003c/li\u003e\n\u003cli\u003eGomes HM, Barddal JP, Enembreck F, Bifet A (2017) A survey on ensemble learning for data stream classification. ACM Comput Surv (CSUR) 50:1-36.\u003c/li\u003e\n\u003cli\u003eGonz\u0026aacute;lez S, Garc\u0026iacute;a S, Del Ser J, Rokach L, Herrera F (2020) A practical tutorial on bagging and boosting based ensembles for machine learning: Algorithms, software tools, performance study, practical perspectives and opportunities. Inform Fusion 64:205-237. https://doi.org/10.1016/j.inffus.2020.07.007\u003c/li\u003e\n\u003cli\u003eGourabi B R, Rad TG (2013) The analysis of ecotourism potential in Boujagh wetland with AHP method. Life Science Journal 10: 251-258.\u003c/li\u003e\n\u003cli\u003eHamze-Ziabari S M, Bakhshpoori T (2018) Improving the prediction of ground motion parameters based on an efficient bagging ensemble model of M5\u0026prime; and CART algorithms. Appl Soft Comput, 68:147-161. https://doi.org/10.1016/j.asoc.2018.03.052\u003c/li\u003e\n\u003cli\u003eHoang HT, Truong QH, Nguyen AT, Hens L (2018) Multicriteria evaluation of tourism potential in the central highlands of Vietnam: Combining geographic information system (GIS), analytic hierarchy process (AHP) and principal component analysis (PCA). Sustainability 10:3097. https://doi.org/10.3390/su10093097\u003c/li\u003e\n\u003cli\u003eHolloway JC, Humphreys C (2022) The Business of Tourism. SAGE.\u003c/li\u003e\n\u003cli\u003eHothorn T, Hornik K, Zeileis, A (2015) ctree: Conditional inference trees. The comprehensive R archive network 8: 1-33.\u003c/li\u003e\n\u003cli\u003ehttps://apfbcs.nic.in/apfbcs/wetland/annexure2.pdf\u003c/li\u003e\n\u003cli\u003ehttps://tourism.assam.gov.in/portlets\u003c/li\u003e\n\u003cli\u003eHuff C, Tingley D (2015) Who are these people? Evaluating the demographic characteristics and political preferences of MTurk survey respondents. Research \u0026amp; Politics 2: 2053168015604648. https://doi.org/10.1177/2053168015604648.\u003c/li\u003e\n\u003cli\u003eHuismann W (2014) PandaLeaks: The Dark Side of the WWF. Nordbook.\u003c/li\u003e\n\u003cli\u003eIslam MT, Ayon EH, Ghosh BP, MD SC, Shahid R, Rahman S, ... Nguyen TN (2024) Revolutionizing Retail: A Hybrid Machine Learning Approach for Precision Demand Forecasting and Strategic Decision-Making in Global Commerce. Journal of Computer Science and Technology Studies 6:33-39. https://doi.org/10.32996/jcsts.2024.6.1.4\u003c/li\u003e\n\u003cli\u003eKachniewska MA (2015) Tourism development as a determinant of quality of life in rural areas. Worldwide Hospitality and Tourism Themes 7:500\u0026ndash;515. https://doi.org/10.1108/WHATT-06-2015-0028\u003c/li\u003e\n\u003cli\u003eKarakitsiou A, Mavrommati A (2017) Machine learning methods in tourism demand forecasting: Some evidence from Greece. MIBES transactions 11:92-105. https://www.researchgate.net/publication/323427035_Machine_learning_methods_in_tourism_demand_forecasting_some_evidence_from_Greece\u003c/li\u003e\n\u003cli\u003eKarali A, Das S, Roy H (2021) Forty years of the rural tourism research: Reviewing the trend, pattern and future agenda. Tourism Recreation Research 49:173-200. https://doi.org/10.1080/02508281.2021.1961065.\u003c/li\u003e\n\u003cli\u003eKatelieva M, Muhar A (2022) Heritage tourism products based on traditional nature-related knowledge: Assessment of cultural, social, and environmental factors in cases from rural Austria. Journal of Heritage Tourism 17:631\u0026ndash;647. https://doi.org/10.1080/1743873X.2022.2098040\u003c/li\u003e\n\u003cli\u003eKhadka D, Chaudhary A, Karki R, Sharma B, Bhatta S (2021) Ecotourism in Ghoda Ghodi Wetland Sukhad, Kailali, Nepal. Journal of Tourism and Hospitality Education 11:22\u0026ndash;42. https://doi.org/10.3126/jthe.v11i0.38237\u003c/li\u003e\n\u003cli\u003eKontogeorgopoulos N (2017) Finding oneself while discovering others: An existential perspective on volunteer tourism in Thailand. Ann Tourism Res 65:1\u0026ndash;12. https://doi.org/10.1016/j.annals.2017.04.006\u003c/li\u003e\n\u003cli\u003eKuhn, Johnson (2013) Applied Predictive Modeling, Springer, New York (chapter 4).\u003c/li\u003e\n\u003cli\u003eLaw R, Li G, Fong DKC, Han X (2019) Tourism demand forecasting: A deep learning approach. Ann Tourism Res 75:410-423. https://doi.org/10.1016/j.annals.2019.01.014\u003c/li\u003e\n\u003cli\u003eLe TH, Arcodia C, Novais MA, Kralj A (2021) Proposing a systematic approach for integrating traditional research methods into machine learning in text analytics in tourism and hospitality. Curr Issues Tour 24:1640-1655. https://doi.org/10.1080/13683500.2020.1829568\u003c/li\u003e\n\u003cli\u003eLinardos V, Drakaki M, Tzionas P, Karnavas YL (2022) Machine learning in disaster management: recent developments in methods and applications. Machine Learning and Knowledge Extraction, 4:446-473. https://doi.org/10.3390/make4020020\u003c/li\u003e\n\u003cli\u003eLiu Z, Peng C, Work T, Candau JN., DesRochers A. Kneeshaw D (2018) Application of machine-learning methods in forest ecology: recent progress and future challenges. Environ Rev 26: 339-350. https://doi.org/10.1139/er-2018-0034\u003c/li\u003e\n\u003cli\u003eLu H, Karimireddy SP, Ponomareva N, Mirrokni V (2020) Accelerating gradient boosting machines. In International Conference on Artificial Intelligence and Statistics :516-526. PMLR.\u003c/li\u003e\n\u003cli\u003eManzoor F, Wei L, Asif M, Haq MZ ul, Rehman H ur. (2019) The Contribution of Sustainable Tourism to Economic Growth and Employment in Pakistan. Int J Environ Res Pub He 16:3785. https://doi.org/10.3390/ijerph16193785.\u003c/li\u003e\n\u003cli\u003eMar\u0026iacute;n-Buz\u0026oacute;n C, P\u0026eacute;rez-Romero A, L\u0026oacute;pez-Castro J L, Ben Jerbania I, Manzano-Agugliaro F (2021) Photogrammetry as a new scientific tool in archaeology: Worldwide research trends. Sustainability 13:5319. https://doi.org/10.3390/su13095319\u003c/li\u003e\n\u003cli\u003eMasselink R, Temme A JAM, Gim\u0026eacute;nez R, Casal\u0026iacute; J, Keesstra SD (2017) Assessing hillslope-channel connectivity in an agricultural catchment using rare-earth oxide tracers and random forests models. Cuadernos De Investigaci\u0026oacute;n Geogr\u0026aacute;fica 43:17\u0026ndash;39. https://doi.org/10.18172/cig.3169\u003c/li\u003e\n\u003cli\u003eMemon MA, Cheah JH, Ramayah T, Ting H, Chuah F, Cham TH. (2019) Moderation analysis: issues and guidelines. Journal of Applied Structural Equation Modeling 3: 1-11.\u003c/li\u003e\n\u003cli\u003eMogensen UB, Ishwaran H, Gerds TA (2012) Evaluating random forests for survival analysis using prediction error curves. J Stat Softw 50:1. https://doi.org/10.18637%2Fjss.v050.i11\u003c/li\u003e\n\u003cli\u003eMohammed A, Kora R (2023) A comprehensive review on ensemble deep learning: Opportunities and challenges. Journal of King Saud University-Computer and Information Sciences 35:757-774.\u003c/li\u003e\n\u003cli\u003eMosebo Fernandes AC, Quintero Gonzalez R, Lenihan-Clarke MA, Leslie Trotter EF, Jokar Arsanjani J (2020) Machine learning for conservation planning in a changing climate. Sustainability 12: 7657. https://doi.org/10.3390/su12187657\u003c/li\u003e\n\u003cli\u003eMienye ID, Sun Y (2022) A survey of ensemble learning: Concepts, algorithms, applications, and prospects. IEEE Access 10:99129-99149.\u003c/li\u003e\n\u003cli\u003eMitra R, Saha P, Das J (2022) Assessment of the performance of GIS-based analytical hierarchical process (AHP) approach for flood modelling in Uttar Dinajpur district of West Bengal, India. Geomatics, Natural Hazards and Risk 13:2183-2226. https://doi.org/10.1080/19475705.2022.2112094\u003c/li\u003e\n\u003cli\u003eMunier, N., Hontoria, E. (2021) Uses and Limitations of the AHP Method. Cham: Springer International Publishing.\u003c/li\u003e\n\u003cli\u003eNaghibi SA, Ahmadi K, Daneshi A (2017) Application of support vector machine, random forest, and genetic algorithm optimized random forest models in groundwater potential mapping. Water Resour Manag 31:2761-2775. https://doi.org/10.1007/s11269-017-1660-3\u003c/li\u003e\n\u003cli\u003eNath B, Wang Z, Ge Y, Islam K, P Singh, R., Niu, Z (2020) Land use and land cover change modeling and future potential landscape risk assessment using Markov-CA model and analytical hierarchy process. ISPRS Int J Geo-Inf 9:134. https://doi.org/10.3390/ijgi9020134\u003c/li\u003e\n\u003cli\u003eOpitz D, Maclin R (1999) Popular ensemble methods: An empirical study. J Artif Intel Res. 11: 169\u0026ndash;198.\u003c/li\u003e\n\u003cli\u003ePathmanandakumar V, Goh HC, Chenoli SN (2023) Identifying potential zones for ecotourism development in Batticaloa district of Sri Lanka using the gis-based Ahp spatial analysis. GeoJournal of Tourism and Geosites 46:252-261. https://doi.org/10.30892/gtg.46128-1022\u003c/li\u003e\n\u003cli\u003ePolikar R (2006) Ensemble based systems in decision making. IEEE Circ Syst Mag 6:21-45.\u003c/li\u003e\n\u003cli\u003ePourghasemi HR, Rahmati O (2018) Prediction of the landslide susceptibility: Which algorithm, which precision? Catena 162:177-192.\u003c/li\u003e\n\u003cli\u003ePuh K, Bagić Babac M (2023) Predicting sentiment and rating of tourist reviews using machine learning. Journal of hospitality and tourism insights 6:1188-1204. https://doi.org/10.1108/JHTI-02-2022-0078\u003c/li\u003e\n\u003cli\u003ePu\u0026scaron;ka A, Pamucar D, Stojanović I, Cavallaro F, Kaklauskas A, Mardani A (2021) Examination of the Sustainable Rural Tourism Potential of the Brčko District of Bosnia and Herzegovina Using a Fuzzy Approach Based on Group Decision Making. Sustainability 13: 583. https://doi.org/10.3390/su13020583\u003c/li\u003e\n\u003cli\u003eQuinlan JR (1992) Learning with continuous classes. In 5th Australian joint conference on artificial intelligence 92:343-348).\u003c/li\u003e\n\u003cli\u003eRaha S, Gayen SK (2023) Application of Analytic Hierarchy Process and weighted sum techniques for green tourism potential mapping in the Gangetic West Bengal, India. GeoJournal 88(Suppl 1): 197-240. https://doi.org/10.1007/s10708-022-10619-2\u003c/li\u003e\n\u003cli\u003eRaha S, Gayen SK (2021) Tourism Potentiality Zone Mapping by Using the AHP Technique: A Study on Bankura District, West Bengal, India. Journal of Geographical Studies\u003cem\u003e \u003c/em\u003e6: 58-85. https://doi.org/10.21523/gcj5.22060201\u003c/li\u003e\n\u003cli\u003eRaha S, Mondal M, Gayen SK (2021) Ecotourism potential zone mapping by using analytic hierarchy process (AHP) and weighted linear algorithm: A study on West Bengal, India. Journal of Geographical Studies 5: 44-64. https://doi.org/10.21523/gcj5.21050201\u003c/li\u003e\n\u003cli\u003eRidgeway G (2007) Generalized Boosted Models: A guide to the gbm package. Update, 1: 2007.\u003c/li\u003e\n\u003cli\u003eRincy TN, Gupta R (2020) Ensemble learning techniques and its efficiency in machine learning: A survey. In 2nd international conference on data, engineering and applications (IDEA) IEEE pp:1-6.\u003c/li\u003e\n\u003cli\u003eRokach L (2010) Ensemble-based classifiers. Artificial Intelligence Review 33: 1\u0026ndash;39\u003c/li\u003e\n\u003cli\u003eRonghang S, Sen Dr S (2022) Tourism potentials in the Karbi Anglong autonomous council districts (KAAC) of Assam. International Journal of Geography, Geology and Environment 4:112\u0026ndash;117. https://doi.org/10.22271/27067483.2022.v4.i2b.123.\u003c/li\u003e\n\u003cli\u003eRoodposhti MS, Safarrad T, Shahabi H (2017) Drought sensitivity mapping using two one-class support vector machine algorithms. Atmos Res 193:73-82. https://doi.org/10.1016/j.atmosres.2017.04.017\u003c/li\u003e\n\u003cli\u003eRoy SK, Hasan MM, Mondal I, Akhter J, Roy SK, Talukder S, ... Karuppannan S (2024) Empowered machine learning algorithm to identify sustainable groundwater potential zone map in Jashore District, Bangladesh. Groundwater for Sustainable Development 25: 101168. https://doi.org/10.1016/j.gsd.2024.101168\u003c/li\u003e\n\u003cli\u003eSachdeva S, Kumar B (2021) Comparison of gradient boosted decision trees and random forest for groundwater potential mapping in Dholpur (Rajasthan), India. Stoc Env Res Risk A 35:287-306. https://doi.org/10.1007/s00477-020-01891-0\u003c/li\u003e\n\u003cli\u003eSahani N (2020) Application of analytical hierarchy process and GIS for ecotourism potentiality mapping in Kullu District, Himachal Pradesh, India. Environment, Development and Sustainability 22: 6187-6211. https://doi.org/10.1007/s10668-019-00470-w\u003c/li\u003e\n\u003cli\u003eSaaty TL (1980) The analytic hierarchy process: Planning, priority setting, resource allocation. McGraw Hill.\u003c/li\u003e\n\u003cli\u003eSaaty, RW (1987) The analytic hierarchy process\u0026mdash;What it is and how it is used. Math Modelling 9:161\u0026ndash;176. https://doi.org/10.1016/0270-0255(87)90473-8.\u003c/li\u003e\n\u003cli\u003eScarpocchi C (2020) Where are people going? In \u003cem\u003ePlace Branding\u003c/em\u003e. Routledge.\u003c/li\u003e\n\u003cli\u003eSchultze, J, G\u0026auml;rtner S, Bauhus J, Meyer P, Reif A (2014) Criteria to evaluate the conservation value of strictly protected forest reserves in Central Europe. Biodivers Conserv 23:3519-3542.\u003c/li\u003e\n\u003cli\u003eSenapati U, Das TK (2021) Assessment of basin-scale groundwater potentiality mapping in drought-prone upper Dwarakeshwar River basin, West Bengal, India, using GIS-based AHP techniques. Arab J Geosci 14:960. https://doi.org/10.1007/s12517-021-07316-8\u003c/li\u003e\n\u003cli\u003eSingh KR, Goswami AP, Kalamdhad AS, Kumar B (2020) Assessment of surface water quality of Pagladia, Beki and Kolong rivers (Assam, India) using multivariate statistical techniques. International Journal of River Basin Management 18: 511\u0026ndash;520. https://doi.org/10.1080/15715124.2019.1566236\u003c/li\u003e\n\u003cli\u003eSingh S, Rao MJ, Baranval NK, Kumar, KV, Kumar YV (2023) Geoenvironment factors guided coastal urban growth prospect (UGP) delineation using heuristic and machine learning models. Ocean Coast Manage. 236:106496. https://doi.org/10.1016/j.ocecoaman.2023.106496\u003c/li\u003e\n\u003cli\u003eStrobl C, Boulesteix AL, Zeileis A, Hothorn T (2007) Bias in random forest variable importance measures: Illustrations, sources and a solution. BMC Bioinformatics 8:1-21.\u003c/li\u003e\n\u003cli\u003eSun S, Wei Y, Tsui KL, Wang S (2019) Forecasting tourist arrivals with machine learning and internet search index. Tourism Manage 70:1-10. https://doi.org/10.1016/j.tourman.2018.07.010\u003c/li\u003e\n\u003cli\u003eTaalab K, Cheng T, Zhang Y (2018) Mapping landslide susceptibility and types using Random Forest. Big Earth Data 2:159-178.\u003c/li\u003e\n\u003cli\u003eTang Z, Zhao W, Xie X, Zhong Z, Shi F, Liu J, Shen D (2020) Severity assessment of coronavirus disease 2019 (COVID-19) using quantitative features from chest CT images. arXiv preprint arXiv:2003.11988. https://doi.org/10.48550/arXiv.2003.11988\u003c/li\u003e\n\u003cli\u003eTekouabou SCK, Diop EB, Azmi R, Jaligot R, Chenal J (2022) Reviewing the application of machine learning methods to model urban form indicators in planning decision support systems: Potential, issues and challenges. Journal of King Saud University-Computer and Information Sciences 34: 5943-5967. https://doi.org/10.1016/j.jksuci.2021.08.007\u003c/li\u003e\n\u003cli\u003eThanh NN, Chotpantarat S, Trung NH., Ngu NH (2022) Mapping groundwater potential zones in Kanchanaburi Province, Thailand by integrating of analytic hierarchy process, frequency ratio, and random forest. Ecol Indic 145:109591. https://doi.org/10.1016/j.ecolind.2022.109591\u003c/li\u003e\n\u003cli\u003eTrabelsi, F, Bel Hadj Ali S, Lee S (2023) Comparison of Novel Hybrid and Benchmark Machine Learning Algorithms to Predict Groundwater Potentiality: Case of a Drought-Prone Region of Medjerda Basin, Northern Tunisia. Remote Sens Basel 15:152. https://doi.org/10.3390/rs15010152\u003c/li\u003e\n\u003cli\u003eTrukhachev A (2015) Methodology for evaluating the rural tourism potentials: A tool to ensure sustainable development of rural settlements. Sustainability 7: 3052-3070.\u003c/li\u003e\n\u003cli\u003eYoussef AM, Pourghasemi HR, Pourtaghi ZS, Al-Katheeri MM (2016) Erratum to: Landslide susceptibility mapping using random forest, boosted regression tree, classification and regression tree, and general linear models and comparison of their performance at Wadi Tayyah Basin, Asir Region, Saudi Arabia. Landslides, 13, 1315-1318.\u003c/li\u003e\n\u003cli\u003eVabalas A, Gowen E, Poliakoff E, Casson AJ (2019) Machine learning algorithm validation with a limited sample size. PloS one 14: e0224365. https://doi.org/10.1371/journal.pone.0224365\u003c/li\u003e\n\u003cli\u003eVafadar S, Rahimzadegan M, Asadi R (2023) Evaluating the performance of machine learning methods and Geographic Information System (GIS) in identifying groundwater potential zones in Tehran-Karaj plain, Iran. J Hydrol 624:129952. https://doi.org/10.1016/j.jhydrol.2023.129952\u003c/li\u003e\n\u003cli\u003eVrontos SD, Galakis J, Vrontos ID. (2021) Modeling and predicting US recessions using machine learning techniques. Int J Forecasting 37: 647-671. https://doi.org/10.1016/j.ijforecast.2020.08.005\u003c/li\u003e\n\u003cli\u003eWang C, Yu Q, Law KH, McKenna F, Stella XY., Taciroglu E, ... Cetiner B (2021) Machine learning-based regional scale intelligent modeling of building information for natural hazard risk management. Automat Constr, 122:103474. https://doi.org/10.1016/j.autcon.2020.103474\u003c/li\u003e\n\u003cli\u003eYuxi, Z, Linsheng, Z. (2020) Identifying conflicts tendency between nature-based tourism development and ecological protection in China. Ecol Indic, 109:105791.\u003c/li\u003e\n\u003cli\u003eZabihi H, Alizadeh M, Wolf ID, Karami M, Ahmad A, Salamian H (2020) A GIS-based fuzzy-analytic hierarchy process (F-AHP) for ecotourism suitability decision making: A case study of Babol in Iran. Tourism Management Perspectives 36:100726. https://doi.org/10.1016/j.tmp.2020.100726\u003c/li\u003e\n\u003cli\u003eZekan B, Weismayer C, Gunter U, Schuh B, Sedlacek S (2022) Regional sustainability and tourism carrying capacities. J Clean Prod 339:130624. https://doi.org/10.1016/j.jclepro.2022.130624\u003c/li\u003e\n\u003cli\u003eZhang K, Wu X, Niu R, Yang K, Zhao L (2017) The assessment of landslide susceptibility mapping using random forest and decision tree methods in the Three Gorges Reservoir area, China. Environ Earth Sci 76:1-20.\u003c/li\u003e\n\u003cli\u003eZhao Q, Yu S, Zhao F, Tian L, Zhao Z (2019) Comparison of machine learning algorithms for forest parameter estimations and application for forest quality assessments. Forest Ecol Manag 434:224-234. https://doi.org/10.1016/j.foreco.2018.12.019.\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"Tourism potentiality, Analytic Hierarchy Process, ROC-AUC, Conditional Inference Tree, Bagged CART, Random Forest, TPZ Ensemble model","lastPublishedDoi":"10.21203/rs.3.rs-4364952/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-4364952/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eAlthough Assam is enriched with several popular tourist destinations but till date, its\u0026rsquo; complete charm remains enigmatic. This research was aimed at prognosticating the Tourism Potential Zone (TPZ) for the state of Assam using five machine learning (i.e., Conditional Inference Tree, Bagged CART, Random Forest, Random Forest with Conditional Inference Tree, and Gradient Boosting models) and one ensemble model. A 5-step methodology was implemented to do this research. First, a Tourism Inventory Database was prepared using the Google earth Imagery, and a rapid field investigation carried out with the help of Global Positioning System and non-participant observation technique. Total 365 tourism points was in the inventory, 70% (224) of which was used for the training set and 30% (124) was used for the validation purpose. The tourism conditioning factors such as Relief, Aspect, Viewshed, Forest Area, Wetland, Coefficient of Variation of rainfall, Reserve Forest, Population Density, Population Growth Rate, Literacy Rate and Road-railway density were used as the independent variables in the modelling process. The TPZ was predicted with the help of above machine learning models and finally, a new TPZ Ensemble Model was proposed by combining each model. The result showed that all machine learning models performed well according to prediction accuracy and finally, the ensemble model outperformed other models by achieving the highest AUC (97.6%), Kappa (0.82) and accuracy (0.93) values. The results obtained from this research using machine learning and ensemble methods can provide proper and significant information for decision makers for the development of tourism in the region.\u003c/p\u003e","manuscriptTitle":"Harnessing Machine Learning and Ensemble Models for Tourism Potential Zone Prediction for the Assam State of India","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2024-05-07 16:55:44","doi":"10.21203/rs.3.rs-4364952/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"85ec27c1-5464-434d-bd9b-8eb2ffce0f65","owner":[],"postedDate":"May 7th, 2024","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":31489753,"name":"Artificial Intelligence and Machine Learning"},{"id":31489754,"name":"Hospitality and Tourism"}],"tags":[],"updatedAt":"2024-05-07T16:55:44+00:00","versionOfRecord":[],"versionCreatedAt":"2024-05-07 16:55:44","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-4364952","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-4364952","identity":"rs-4364952","version":["v1"]},"buildId":"-HB7Z8yhvgn0wM9Nzuekk","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}