Towards Safer Water: AI-Driven Predictive Analytics for Disease Detection

Towards Safer Water: AI-Driven Predictive Analytics for Disease Detection | 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 Case Report Towards Safer Water: AI-Driven Predictive Analytics for Disease Detection Jaya Zalte, Harshal shah This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-6338765/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 Water quality is a critical factor for human health and environmental sustainability. Rapid urbanization and industrialization have led to significant water contamination, increasing the prevalence of waterborne diseases. This study investigates the presence of pathogens in water sources across the Gujarat region, utilizing machine learning models to analyze contamination patterns. Various classifiers, including HistGradientBoosting, Random Forest, AdaBoost, Bagging, Decision Tree, and LSTM, were employed to predict water quality and identify pathogens. Among these, the Random Forest and Bagging classifiers exhibited the highest accuracy at 98.53%. Furthermore, Explainable AI techniques, specifically SHapley Additive exPlanations (SHAP), were used to interpret the significant features influencing contamination levels. The study highlights the need for proactive water quality monitoring and pathogen detection to prevent disease outbreaks. Environmental Engineering Health Economics & Outcomes Research Marine and Freshwater Ecology Machine Learning Water Quality SHAP Explainable AI Contaminants Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 1. Introduction Water is an essential resource for life, yet its quality is often compromised by contamination from domestic, industrial, and agricultural activities. Gujarat, like many regions in India, faces challenges in ensuring safe drinking water due to high dependence on groundwater and inadequate waste management. According to government statistics, Gujarat has approximately 50 billion cubic meters (BCM) of water, with over 80% allocated to irrigation. The remaining supply, often compromised by pollutants and pathogens, poses a significant health risk. Waterborne diseases are a major public health concern. The World Health Organization (WHO) estimates that 7.3 million deaths occur annually due to diarrheal diseases, with children being the most affected. Lack of awareness and sanitation exacerbates the spread of infections. Traditional water quality assessments focus on detecting common contaminants; however, rare pathogens responsible for severe diseases are often overlooked. For instance, a recent outbreak of Guillain-Barré syndrome (GBS) in Pune was linked to the pathogen Campylobacter jejuni, present in contaminated water. This highlights the necessity of advanced analytical approaches for early pathogen detection. This study employs machine learning models to analyze water quality data collected from Gujarat, identifying potential pathogens and assessing contamination levels. Our contributions to this study make an early warning signs for the use of potable water for common people. By integrating Explainable AI techniques, we aim to provide transparency in model predictions, ensuring actionable insights for policymakers and public health officials. 1.1 Related Study This section provides an overview of relevant studies on water quality assessment and machine learning-based prediction models. 1.1.2 Traditional and Statistical Approaches to Water Quality Assessment Multivariate statistical analysis was applied in a Brazilian river pilot study, identifying pollution as a major environmental threat in specific areas [ 1 ]. A two-year study further analyzed water quality parameters and 42 pesticides to assess contamination levels [ 2 ]. Non-machine learning approaches have also been explored, such as various methodologies for Water Quality Index (WQI) calculation [ 6 ]. A study in Tonle Sap Lake, Sangker River (Cambodia) compared five water quality assessment techniques, including the Mekong River Commission WQI, the French Water Quality Assessment, and the US Environmental Protection Agency framework [ 7 ]. 1.1.3 Hybrid Models and Advanced AI Techniques Machine learning techniques have been widely applied in water quality assessments. An ensemble learning model was used for water quality classification [ 3 ], while a study in Lam Tsuen River, Hong Kong, utilized the WQI with an Extra Tree regression model [ 3 ]. Another study employed an XGBoost classifier, achieving 97.06% accuracy using hyperparameter optimization [ 4 ]. Several hybrid models have been explored to enhance predictive accuracy. A study in the Talar catchment found that a Bagging classifier with a Random Tree outperformed other machine learning models for river quality prediction [ 5 ]. Comparative studies have evaluated multiple classifiers, including Support Vector Machine (SVM), Random Forest (RF), Logistic Regression (LR), Decision Tree (DT), CATBoost, XGBoost, and Multilayer Perceptron (MLP) [ 10 ]. The CATBoost model demonstrated the highest accuracy, with feature importance aligning with key pollution indicators [ 10 ]. Spearman rank correlation coefficients were further used to determine significant trends in pollution indicators [ 11 ]. To facilitate early detection of water contamination, researchers in [ 12 ] developed a predictive model for BOD₅ values, using linear regression, support vector regression (SVR), and multi-layer perceptrons (MLP) [ 13 ]. Additionally, various chemometric techniques—such as Man-Kendall trend analysis, principal component analysis, factor analysis, and agglomerative hierarchical cluster analysis—have demonstrated the vulnerability of Selangor's water system to ammonia pollutants, posing significant risks to water supply [ 14 ]. 1.1.4 The Impact of External Factors on Water Quality The impact of COVID-19 on the Ganges River (India) has been studied extensively [ 15 ]. Various WQI methodologies and water quality parameters have been examined in detail [ 16 – 17 ]. In [ 18 ]-[ 19 ], Cascaded Fuzzy Systems were implemented for water quality prediction, while [20] utilized KNN imputation with 10 water quality parameters. 1.1.5 Explainable AI in Water Quality Assessment The integration of Explainable Artificial Intelligence (XAI) in water quality assessment has gained attention in recent years. SHAP (SHapley Additive exPlanations) has been employed to interpret machine learning models, providing transparency in water quality prediction [ 21 – 25 ]. The use of XAI ensures that AI-driven decisions are interpretable and actionable for stakeholders, improving trust in automated water quality assessments. With traditional Water Quality Index (WQI) value calculation, many errors are introduced; to overcome such errors, we can use machine learning and deep learning models. 1.2 Research Gap Despite significant advancements in water quality assessment and prediction models, several key gaps remain in the existing literature: 1. Lack of Explainability in AI-Based Water Quality Predictions While machine learning models such as XGBoost, Random Forest, and neural networks have shown high accuracy in water quality prediction [ 3 , 4 , 10 ], they often operate as black-box models, making it difficult to understand how predictions are made. Few studies have incorporated Explainable AI (XAI) techniques like SHAP (SHapley Additive ExPlanations) to provide insights into model decisions [ 21 – 25 ]. However, existing XAI applications in water quality assessment remain limited and underexplored. 2. Insufficient Hybrid and Ensemble Learning Approaches Although individual models such as XGBoost and Support Vector Machines (SVM) have demonstrated promising results [ 4 , 10 , 13 ], hybrid models and ensemble learning techniques remain underutilized. Studies have shown that combining multiple models, such as Bagging classifiers and Random Trees [ 5 ], can enhance predictive performance, but comprehensive evaluations across different water bodies and geographic locations are lacking. 3. Limited Geographic and Environmental Coverage Most water quality studies have focused on specific rivers and regions, such as: Lam Tsuen River, Hong Kong [ 3 ] Talar Catchment, Iran [ 5 ] Ganges River, India (COVID-19 impact) [ 15 ] Selangor, Malaysia (chemometric analysis) [ 14 ] However, comprehensive datasets covering diverse hydrological conditions and climatic variations are scarce. There is a need for models that generalize across multiple regions and water sources, including groundwater, lakes, and reservoirs. 4. Lack of Real-Time and Early Warning Systems Most studies rely on historical data for water quality prediction [ 12 , 13 , 14 ], rather than real-time monitoring and early warning systems. The integration of IoT sensors, remote sensing, and AI-driven alert mechanisms for water contamination detection remains an open research area. 5. Insufficient Studies on AI-Driven Decision Support Systems While various models assess water quality, few studies explore how AI-based predictions can be effectively integrated into policymaking and public health strategies. There is a lack of user-friendly AI-driven decision support systems that can help environmental agencies and policymakers take preventive actions before contamination reaches critical levels. 6. Data Imbalance and Feature Selection Challenges Many studies face imbalanced datasets, where instances of extreme pollution events are rare [20]. Furthermore, feature selection methodologies are often inconsistent, leading to suboptimal model performance. There is a need for automated feature selection techniques and strategies to handle data imbalance for more robust AI models. Motivation and Contribution This research presents a novel approach for waterborne disease detection by integrating efficient machine learning models with Explainable AI (XAI). The key contributions of this study include: Implementation of advanced machine learning techniques (e.g., XGBoost, Random Forest, and ensemble learning) to classify and predict waterborne disease susceptibility based on water quality parameters.Comparison of various machine learning models to identify the most accurate and efficient approach for disease prediction. Utilization of SHAP (SHapley Additive ExPlanations) to interpret feature importance, ensuring transparency in water quality assessments and disease prediction outcomes. Identification of the most influential water quality parameters contributing to contamination and disease risk, aiding in better decision-making. Application of data balancing techniques to handle class imbalances in water quality datasets, improving model robustness.Feature selection and engineering to enhance predictive performance by reducing noise and redundancy in data. Performance benchmarking against existing models that use traditional Water Quality Index (WQI) and non-explainable ML approaches. Demonstration of how XAI improves interpretability and decision support compared to black-box models. Proposal of an AI-driven decision support system to assist environmental agencies and policymakers in early disease outbreak detection. Contribution : A pilot study was conducted in western parts of Gujarat, Vadodara an analysis of a number of open and closed gutters in the area was surveyed. Along with 250 people were surveyed for notable diseases related to water were identified and analyzed. As per the Fig. 1 , study conducted in the western region of India in Gujarat shows the count of disposing the waste in open gutter is more as compared other waste water management areas. The open gutters are an open invitation for water contamination. Figure 1 ., shows the number of open gutters is more in Gujarat's western region, leading to contamination of water and making it more polluted. Also, the sewage lines in most of the areas remain open, and disposal of wastewater from industries and households leads to severe waterborne diseases. Around 250 people were surveyed for various diseases affected by drinking water or using water repositories near their houses. The Fig. 2 . shows us the identified signs and symptoms observed from 250 people who were also surveyed for drinking habits, water reservoirs nearby, and Diarrheal diseases, which were reported and are plotted in the graph. Table 1 Diseases caused by associated pathogens Diseases Bacteria Health Significance Virus Health Significance Parasites Health Significance Diarrhea Escherichia coli (E. coli) High Rotavirus High Cryptosporidium High Campylobacter jejuni High Norovirus High Giardia High Salmonella High Adenovirus Moderate Entamoeba High Shigella High Astrovirus Moderate -- -- Yersinia Moderate -- -- Cholera Vibrio cholerae High -- -- -- Hepatitis A hepatitis A virus (HAV) High -- -- Typhoid Salmonella Typhi bacteria High -- -- Giardiasis Giardia duodenalis. Giardia intestinalis. High Pnemonia , Meningitis , Urinary Tract infection Enterohemorrhagic E. coli (EHEC) High -- -- -- --- Campylobacteriosis Guillain-Barré Syndrome • Campylobacter jejuni. • Campylobacter coli High -- -- -- --- Cryptosporidiosis --- --- -- -- Cryptosporidium High Cyclosporiasis --- ---- -- -- Cyclospora cayetanensis High Legionellosis Legionella High -- -- ---- -- Shigellosis Shigella bacteria High -- -- ---- -- Vibriosis Vibrio parahaemolyticus and Vibrio vulnificus. High -- -- ---- -- Table. 1, describes the diseases associated with the relevant pathogen categorized as Bacteria, Virus and Parasites. Health significance column, relates to the severity of impact with low, moderate and high values, including association with outbreaks. 2. Materials and Methods The dataset is collected from Pollution Board Control consists of more than 2700 data, for 5 years ranging from 2017 to 2022. Each attribute in the dataset consists of the water parameters like pH, BOD, Dissolved solids, Temperature, Conductivity, Nitrogen, Fecal Coliform, and Fecal Streptococci. The dataset consists of river water from various states of India. Overall, 22 states of India were taken into account. All this data was given to the machine learning model for the training phase of the model. The data is pre-processed to be fed into the learning model/ classifiers. This data is cleaned with missing values using mean, median, and mode methodology. The data is used for training and testing used in the model. Data studied and utilized from the pilot study also contains the values that were not detected, those values were also removed and can used for testing purposes in the learning model. 2.1 The Machine learning models HistGradientBoosting Classifier- for Big Data sets of more than 10,000 samples, is much faster as compared to GradientBoostingClassifier. This classifier supports the presence of missing values in the dataset. Consider x and y with two inputs that have samples N. The function f(xi), maps the feature x, which is input to the variable y. The summation is given by the following equation The function is called loss function is given by the difference between the actual and the predicted variables. L(f)=∑_(i = 1)^N [L(yi,f(xi)] (1) Random Forest Classifier - Random Forest supervised machine learning algorithm that combines multiple decision trees to form a forest. Here, GE is the generalization error for the random forest and is denoted as Here, function f (X, Y) is used to count the average of counts and gives predicted value. GE = Px,y ¬(f (X, Y) < 0 (2) Where X is called the predicted value, and Y is called the outcome of the classification problem. Bayesian Classifier - Naive Bayes classifier is a machine learning model based on Bayes' theorem [ 4 ]. It calculates the probability of a given input belonging to a particular class. Here in (3), we have used probabilistic functions to create a classifier model. The probability of given feature inputs for all possible values of the features of y and maximum probability is given as output from the function defined below. y = max(c)Π_(ⅈ=1)^nc(xi∣y) (3) c is called the probability of a particular class, and c(xi∣y) is called conditional probability. Decision tree Classifier- It is a supervised learning technique, hence used for classification and regression for various applications. This Tree classifier has nodes that represent the features of the dataset. Branches indicate the decision rules, and leaves are the outputs of the algorithm. Adaboosting Classifier- This classifier is used to remove the faults that occur in training the model. It is a machine-learning model used for classification and regression problems. Long short-term memory (LSTM) model is used. The LSTM is a deep learning model that retains information for a long series time. 2.2 Algorithm Data preprocessing was done, each of the water quality parameters mentioned had two columns, minimum and maximum. Each feature of water was segregated with minimum values and trained only for those values. Similarly, every feature of maximum value was trained with ML models. Each row was initialized with 0/1 catering to susceptible to diseases. The dataset comprised of missing values, with the SMOTE method, we evaluated the dataset carefully and adjusted the missing value. Each trained model was fed with the processed dataset, and each model gave a fairly good performance. For the Random Forest classifier, Explainable AI was used since this model gave the highest accuracy using the SHapley Additive exPlanations (SHAP) methodology. SHAP is an analytical approach that explains the output of machine learning model. The figure, shows the entire methodology used for models. All the above steps are also depicted in the figure below. 3. Results From the observation of experiments conducted, the classifier described in the methodology section is at par with the dataset consisting of minimum and maximum values. The machine learning algorithms are trained for both minimum and maximum values of water data parameters. Along with the machine learning models, a deep learning model is also trained, which is the long-short-term memory (LSTM) model. The number of layers used is 3, with a dropout value of 0.02. The model used is a sequential model for LSTM. For the LSTM model, the number of epochs used was 15, and we were able to achieve an accuracy of 91.9%. Various classifiers are also deployed for the same data set, which gives us the following observed values. Figure 4 , gives us the brief of all the classifiers used for the model, both values of datasets are considered, minimum values and maximum values. And on both values, all the parameters are trained. The following table shows the trained value accuracy for each of the classifiers used. From the above data, we are also able to calculate the training loss that occurred during the training of the deep learning model. As we can see, deep learning models need huge networked layers, which increases the complexity of the framework. Alone with machine learning model classifiers, we are able to achieve a sufficient amount of accuracy needed to understand the water data. Each of the tests performed is done on random data picked up from the dataset collected from pilot study. 3.1. Performance of the models HistGradientBoosting Classifier is trained on both values of the dataset, minimum range values of water as well as maximum values. Providing a fair accuracy of 97 to 98%. Random Forest outperforms all the other machine learning models. Even though it is one of the traditional methods of evaluating historical data, it gives a good accuracy over the dataset used. Adaboost, Bagging classifier, and Decision Tree provide fairly good performance in terms of maximum values. Since the data is also trained for deep learning models like LSTM to provide a fair comparison with machine learning models, LSTM shows an accuracy of around 91.9%, as most of the deep learning models are used for more complex architectures with many hidden layers. This dataset has a varied range right, from minimum values to maximum values for each of the features. Explainable AI is also called the interpretability of machine learning models used. Machine learning models are thought of as a black box [24–25] . Since the outcome of each model is difficult to trace back to how the results are achieved. Explainable AI helps with the interpretation of each model used. Thus, giving insight into each of the features used and trusting our Machine learning model, which can be safely used over a different range of applications. Since Random Forest is used with XAI, which has outperformed other models, we look forward to getting sustainable results over any other datasets, and we should trust the model in terms of the results obtained. Here, for Explainable AI, we have implemented and installed SHAP (Shapley Additive exPlanations) with the Python framework. Explainable AI is implemented with SHAP for Random Forest Classifier, which provides an accuracy of 98.53%. Feature evaluation is done with SHAP, which provides very good insights about the model trained and various features applied. Figure 5 ., is about the beeswarm model of SHAP, showing high and low values, indicating blue color for low-value impact on the model and red shows high-value impact on the model. Here, we can see temperature shows the high impact on the performance of the Random Forest classifier. In Fig. 6 ., the waterfall diagram, the x-axis highlights the values of the dependent variable, which is susceptible to diseases. x is the observed value, f(x)gives the prediction value of the model for a given input x, and E(x) is the expected value of the dependent variable. The average of all predictions is given by (mean(model(f(x))). We can observe for a certain data value; the Total coliform feature is found to be + 0.58, having more impact as compared to other features in the dataset. Figure 7 ., gives us a detailed picture of the features making an impact on the overall model. Feature importance can be observed from the plot moving toward the base value to the higher or lower side. As observed, the Total coliform has a much higher impact and is larger plot towards the higher side. For lower predictions, BOD and Nitrate and Nitrite values are observed since the impact on the model is on the lower side. The red color represents high, whereas blue shows low values on correlated variables. In this case, the high positive impact of Total Coliform levels shows susceptibility to diseases from the SHAP values lying on the X-axis. Table 2 Comparative table for Existing models and proposed framework Reference Waterbody Location Models used Parameters used Accuracy D. de Andrade Costa,et.al [1] Mixed -- (RF) classifier, decision tree (DT) classifier, (AdaBoost) classifier, (SVM), Naïve Bayes 21 XGBoost 97.06% Bhateria, R., Jain, D [13] Mixed Water bodies India SVM, RF, LR, DT, CATBoost, XGBoost, and MLP [ 13 ]. 7 CATBoost (94.51%) Muduli, P.R., Kumar, A., Kanuri, V.V [15] Water reservoirs Taiwan ANN 6 -- Kambala Vijaya Kumar [19] Water Reservoir Georgia, USA Neuro-Fuzzy (ANFIS) 11 92.37% Azween Abdullah,et.al [20] Kaggle -- Decision Trees, Random Forest, Logistic Regression, Support Vector Machines [20] 21 95.08% Jinal Patel,et.al, [21] Mixed Water -- RF, SVM, Gradient Boost 9 81% Madni HA, et al., [22] Water (Kaggle) -- KNN imputer 10 97% N. Hellen and G. Marvin [23] Mixed water bodies --- RF,ET,KNN,LDA,SVM 10 67% Proposed Work River water India HistGradientBoosting Classifier, RF Forest, AdaBoost Classifier, Bagging Classifier, Decision Tree, LSTM 9 98.53% Table 2 shows the comparative study of the existing frameworks with the proposed framework. As we can see in the table, we can achieve good accuracy as compared to existing Machine learning models. One more point, we have also included minimum and maximum values of each of the parameters, which was missing in all the above-studied papers. 4. Conclusions This study underscores the importance of advanced analytical techniques in water quality assessment and pathogen detection. Through machine learning models, particularly Random Forest and Explainable AI, we successfully identified key contaminants and their impact on public health. The findings emphasize the urgent need for robust water monitoring systems to prevent disease outbreaks and improve water management practices. Future research can extend this work by incorporating microbiological analysis and real-time data integration for predictive water safety measures. By leveraging AI-driven solutions, we can ensure safer water sources and a healthier population. To gain insight into the model trained, we were able to use Explainable AI with the Machine learning model to learn about the features that dominate the parameters leading to waterborne diseases. Further, we can extend the scope by evaluating the microorganisms found in the water. These micro-organisms could be further compared with the signs and symptoms of people identified with the related diseases observed in the study already carried out. To gain insight into the model trained, we were able to use Explainable AI with the Machine learning model to learn about the features that dominate the parameters leading to waterborne diseases. Further, we can extend the scope by evaluating the microorganisms found in the water. These micro-organisms could be further compared with the signs and symptoms of people identified with the related diseases observed in the study already carried out. Declarations Conflict of Interest Authors declare no conflict of interest. Funding Authors state no funding involved. Author Contributions Author -Jaya Zalte, have done cleaning of data, conceptualization, investigation and implementation. Corresponding author- Dr. Harshal shah, have contributed in conceptualization, methodology, inspection and support. Dr. M.H Fulekar have support for the pilot study in Gujarat, he has conceptualised the idea and provided the entire support needed for this project. Acknowledgments I would like to thank, Pollution Board Control of India and Parul university for providing the support. Data Availability The data that support the findings of this study are available on request from the corresponding author, [HS]. The data, which contain information that could compromise the privacy of research participants, are not publicly available due to certain restrictions. References de Andrade Costa D, Soares de Azevedo JP, Dos Santos MA, Dos Santos Facchetti Vinhaes Assumpção R (2020) Water quality assessment based on multivariate statistics and water quality index of a strategic river in the Brazilian Atlantic Forest. Sci Rep 10(1):22038. https://www.nature.com/articles/s41598-020-78563-0 Ramírez-Morales D, Pérez-Villanueva ME, Chin-Pampillo JS, Aguilar-Mora P, Arias-Mora V, Masís-Mora M (2021) Pesticide occurrence and water quality assessment from an agriculturally influenced Latin-American tropical region. Chemosphere. 262,127851 https://doi.org/10.1016/j.chemosphere.2020.127851 Asadollah SBHS, Sharafati A, Motta D, Yaseen ZM (2021) River water quality index prediction and uncertainty analysis: A comparative study of machine learning models. J Environ Chem Eng 9(1) Article 104599. https://doi.org/10.1016/j.jece.2020.104599 Elvin, Wibowo A (2024) Forecasting water quality through machine learning and hyperparameter optimization. JE Indonesian J Electr Eng Comput Sci 496–506. 10.11591/ijeecs.v33.i1.pp496-506 Bui DT, Khosravi K, Tiefenbacher J, Nguyen H, Kazakis N Improving prediction of water quality indices using novel hybrid machine-learning algorithms. Sci Total Environ, 721:137612, 10.1016/j.scitotenv.2020.137612,2020 Tyagi S et al (2018) Water Quality Assessment in Terms of Water Quality Index. Am J Water Resour 13:4–38. 10.11648/j.ajese.20170104.13 Socheat Chrea Loïc, Tudesque R, Chea (2023) Comparative assessment of water quality classification techniques in the largest north-western river of Cambodia (Sangker River-Tonle Sap Basin), Ecological Indicators . 154. https://doi.org/10.1016/j.ecolind.2023.110759 Barzegar Y, Gorelova I, Bellini F, D’Ascenzo F Drinking Water Quality Assessment Using a Fuzzy Inference System Method: A Case Study of Rome (Italy). Int J Environ Res Public Health. ; 20(15):6522, https://doi.org/10.3390/ijerph20156522,2023 Ameen HA Spring water quality assessment using water quality index in villages of Barwari Bala, Duhok, Kurdistan Region, Iraq. Appl Water Sci, 9, https://doi.org/10.1007/s13201-019-1080-z,2019 Nida Nasir A, Kansal O, Alshaltone F, Barneih M, Sameer A, Shanableh Ahmed Al-Shamma'a, Water quality classification using machine learning algorithms. J Water Process Eng, 48, https://doi.org/10.1016/j.jwpe.2022.102920,2022 Yan T, Zhou A, Shen SL (2023) Prediction of long-term water quality using machine learning enhanced by Bayesian optimization. Environ Pollut 318:120870. 10.1016/j.envpol.2022.120870 Jui-Sheng C, Chia-Chun H, Ha-Son H (2018) Determining the quality of water in the reservoir using machine learning. Ecol Inf 44. 10.1016/j.ecoinf.2018.01.005 Bhateria R, Jain D (2016) Water quality assessment of lake water: a review, Sustain. Water Resour Manag 2:161–173. https://doi.org/10.1007/s40899-015-0014-7 Loi JX, Chua ASM, Rabuni MF et al (2022) Water quality assessment and pollution threat to safe water supply for three river basins in Malaysia, Sci Total Environ . ;832:155067. 10.1016/j.scitotenv.2022.155067 , 2022 Muduli PR, Kumar A, Kanuri VV et al Water quality assessment of the Ganges River during COVID-19 lockdown. Int J Environ Sci Technol. 18, 1645–1652, 10.1007/s13762-021-03245-x,2021 Uddin MG, Nash S, Olbert AI A review of water quality index models and their use for assessing surface water quality. Ecol Ind, 122, https://doi.org/10.1016/j.ecolind.2020.107218,2021 Magesh NS, Krishnakumar S, Chandrasekar (2013) Groundwater quality assessment using WQI and GIS techniques. Arab J Geosci. 10.1007/s12517-012-0673-8 Shakour S, Chitsazan M, Mirzaee SY (2023) Zonation of groundwater quality in terms of drinkability, using Fuzzy Logic and Schoeller deterministic method for Northern Dezful - Andimeshk Plain, Iran. Discov Water 3:22. https://doi.org/10.1007/s43832-023-00046-w Kambala V, Kumar Y, Dileep Kumar SR, Godla MSA, Ansari YAB, El-Ebiary, Muniyandy E (2023) Enhancing Water Quality Forecasting Reliability Through Optimal Parameterization of Neuro-Fuzzy Models via Tunicate Swarm Optimization, International Journal of Advanced Computer Science and Applications , 15(3), DOI: 10.14569/IJACSA.2024.01503110,2024. Azween Abdullah, Himakshi Chaturvedi, Siddhesh Fuladi, Nandhika Jhansi Ravuri, Deepa Natesan and M. K Nallakaruppan, Reliable and Efficient Model for Water Quality Prediction and Forecasting, International Journal of Advanced Computer Science and Applications , 14(12), 10.14569/IJACSA.2023.0141219 Patel J, Amipara C, Ahanger TA et al (2022) A Machine Learning-Based Water Potability Prediction Model by Using Synthetic Minority Oversampling Technique and Explainable AI, Comput Intell Neurosci , :9283293. 10.1155/2022/9283293 , 2022 Madni HA, Umer M, Ishaq A, Abuzinadah N, Saidani O, Alsubai S, Hamdi M, Ashraf I (2023) Water-Quality Prediction Based on H 2 O AutoML and Explainable AI Techniques. 15(3):475. https://doi.org/10.3390/w15030475 Nakayiza H, Ggaliwango M (2022) Explainable AI for Safe Water Evaluation for Public Health in Urban Settings, International Conference on Innovations in Science, Engineering and Technology (ICISET), 1–6. https://doi.org/10.1109/iciset54810.2022.9775912 Mallick J, Alqadhi S, Hang HT et al (2024) Interpreting optimised data-driven solution with explainable artificial intelligence (XAI) for water quality assessment for better decision-making in pollution management. Environ Sci Pollut Res 31:42948–42969. https://doi.org/10.1007/s11356-024-33921-7 Randika K, Makumbura L, Mampitiya N, Rathnayake DPP, Meddage S, Henna TL, Dang Y, Hoshino U, Rathnayake Advancing water quality assessment and prediction using machine learning models, coupled with explainable artificial intelligence (XAI) techniques like Shapley additive explanations (SHAP) for interpreting the black-box nature. Results Eng, 23,2024,102831, ISSN 2590 – 1230, https://doi.org/10.1016/j.rineng.2024.102831 Additional Declarations The authors declare no competing interests. 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-6338765","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Case Report","associatedPublications":[],"authors":[{"id":435929934,"identity":"3e5bbac6-e280-4078-ba54-a78027d29c86","order_by":0,"name":"Jaya Zalte","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA+ElEQVRIie2RsUrDUBSGfwk0Dj9kvdJiXuGGQukg9VVuCWSKq4uDmdol0LXvIZSOgYAu1z1jQ6FThltcKgQ0oCIuNx0d7rccOIePcw4/4HD8Qy6yrzoCvAIovtvqDIXEQJ2n/NAplL+KDW+pI4N2xls/fzMnfRPCL3eot5bD8nQswJjk6+Yqr5IoYyIx17ZfUikgPFLcbUBTdl+kwHxhUVbN+AT5SIbN/tiaD4Wg6VHW6URAld0WYsiqUBB9W9aH+6kqXkidTIYjHUcLcZCFTYlW8VNl2odrf1nuj83zLAyCuK7fbUp2Kf8GN0BPPF0MO9vc4XA4HMAnFThM1Mb1a3gAAAAASUVORK5CYII=","orcid":"","institution":"","correspondingAuthor":true,"prefix":"","firstName":"Jaya","middleName":"","lastName":"Zalte","suffix":""},{"id":435929935,"identity":"81eda72b-fd5c-47c0-932f-b8e56ac50e43","order_by":1,"name":"Harshal shah","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Harshal","middleName":"","lastName":"shah","suffix":""}],"badges":[],"createdAt":"2025-03-30 13:46:41","currentVersionCode":1,"declarations":{"humanSubjects":false,"vertebrateSubjects":false,"conflictsOfInterestStatement":false,"humanSubjectEthicalGuidelines":false,"humanSubjectConsent":false,"humanSubjectClinicalTrial":false,"humanSubjectCaseReport":false,"vertebrateSubjectEthicalGuidelines":false},"doi":"10.21203/rs.3.rs-6338765/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-6338765/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":79673803,"identity":"3f516b53-21fc-48ad-9955-49cf504d9fcc","added_by":"auto","created_at":"2025-04-01 11:51:02","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":9956,"visible":true,"origin":"","legend":"\u003cp\u003eDisposal of wastewater\u003c/p\u003e","description":"","filename":"Onlinedrawingimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-6338765/v1/34b52db6f3560bbe07d6fb42.png"},{"id":79673806,"identity":"d98c3c12-dc26-411e-9129-10b85ef8ee66","added_by":"auto","created_at":"2025-04-01 11:51:02","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":7760,"visible":true,"origin":"","legend":"\u003cp\u003eGraph showing the water borne diseases susceptibility to various diseases noted in Gujarat region.\u003c/p\u003e","description":"","filename":"Onlinedrawingimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-6338765/v1/546493181c28f80308a185b6.png"},{"id":79675597,"identity":"766ec8c2-7ecd-47c2-b510-cb6778ef0825","added_by":"auto","created_at":"2025-04-01 12:07:02","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":38085,"visible":true,"origin":"","legend":"\u003cp\u003eMethodology\u003c/p\u003e","description":"","filename":"Onlinefloatimage111.png","url":"https://assets-eu.researchsquare.com/files/rs-6338765/v1/8aed72e221c9e90bc5d0481b.png"},{"id":79675599,"identity":"f5a2da35-904f-42f2-859d-83021242fbbd","added_by":"auto","created_at":"2025-04-01 12:07:02","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":8170,"visible":true,"origin":"","legend":"\u003cp\u003eAccuracy Obtained from Each of The Models Applied on The Maximum and Minimum Values.\u003c/p\u003e","description":"","filename":"Onlinedrawingimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-6338765/v1/c1693e85119b5d0c1e9f87ba.png"},{"id":79674986,"identity":"3dd82774-3830-4d2c-8c3c-20cd70a2727b","added_by":"auto","created_at":"2025-04-01 11:59:02","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":54850,"visible":true,"origin":"","legend":"\u003cp\u003eSHAP values for Beeswarm model\u003c/p\u003e","description":"","filename":"floatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-6338765/v1/3dee631fae2a06cfeb03f743.png"},{"id":79674989,"identity":"83cfb375-8cb5-4def-a040-06c6536e1a69","added_by":"auto","created_at":"2025-04-01 11:59:02","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":36459,"visible":true,"origin":"","legend":"\u003cp\u003eObserved SHAP value for a given f(x)\u003c/p\u003e","description":"","filename":"floatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-6338765/v1/d4b44102bc8a5f57142579a8.png"},{"id":79675598,"identity":"1d95cfa6-923c-4462-a648-1327ad6050f6","added_by":"auto","created_at":"2025-04-01 12:07:02","extension":"jpeg","order_by":7,"title":"Figure 7","display":"","copyAsset":false,"role":"figure","size":33799,"visible":true,"origin":"","legend":"\u003cp\u003eSHAP plot- Feature Importance\u003c/p\u003e","description":"","filename":"floatimage4.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-6338765/v1/1d6be8cfc7a715752faed003.jpeg"},{"id":79676519,"identity":"fe8f49b5-06d8-4f8a-b43e-31007a79daa9","added_by":"auto","created_at":"2025-04-01 12:15:02","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1146445,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6338765/v1/dfd8dd7b-7c41-432d-adcf-e44d3f811aae.pdf"}],"financialInterests":"The authors declare no competing interests.","formattedTitle":"\u003cp\u003eTowards Safer Water: AI-Driven Predictive Analytics for Disease Detection\u003c/p\u003e","fulltext":[{"header":"1. Introduction","content":"\u003cp\u003eWater is an essential resource for life, yet its quality is often compromised by contamination from domestic, industrial, and agricultural activities. Gujarat, like many regions in India, faces challenges in ensuring safe drinking water due to high dependence on groundwater and inadequate waste management. According to government statistics, Gujarat has approximately 50\u0026nbsp;billion cubic meters (BCM) of water, with over 80% allocated to irrigation. The remaining supply, often compromised by pollutants and pathogens, poses a significant health risk.\u003c/p\u003e \u003cp\u003eWaterborne diseases are a major public health concern. The World Health Organization (WHO) estimates that 7.3\u0026nbsp;million deaths occur annually due to diarrheal diseases, with children being the most affected. Lack of awareness and sanitation exacerbates the spread of infections. Traditional water quality assessments focus on detecting common contaminants; however, rare pathogens responsible for severe diseases are often overlooked. For instance, a recent outbreak of Guillain-Barr\u0026eacute; syndrome (GBS) in Pune was linked to the pathogen Campylobacter jejuni, present in contaminated water. This highlights the necessity of advanced analytical approaches for early pathogen detection.\u003c/p\u003e \u003cp\u003eThis study employs machine learning models to analyze water quality data collected from Gujarat, identifying potential pathogens and assessing contamination levels. Our contributions to this study make an early warning signs for the use of potable water for common people. By integrating Explainable AI techniques, we aim to provide transparency in model predictions, ensuring actionable insights for policymakers and public health officials.\u003c/p\u003e \u003cdiv id=\"Sec2\" class=\"Section2\"\u003e \u003ch2\u003e1.1 Related Study\u003c/h2\u003e \u003cp\u003eThis section provides an overview of relevant studies on water quality assessment and machine learning-based prediction models.\u003c/p\u003e \u003cdiv id=\"Sec3\" class=\"Section3\"\u003e \u003ch2\u003e1.1.2 Traditional and Statistical Approaches to Water Quality Assessment\u003c/h2\u003e \u003cp\u003eMultivariate statistical analysis was applied in a Brazilian river pilot study, identifying pollution as a major environmental threat in specific areas [\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e]. A two-year study further analyzed water quality parameters and 42 pesticides to assess contamination levels [\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e]. Non-machine learning approaches have also been explored, such as various methodologies for Water Quality Index (WQI) calculation [\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e]. A study in Tonle Sap Lake, Sangker River (Cambodia) compared five water quality assessment techniques, including the Mekong River Commission WQI, the French Water Quality Assessment, and the US Environmental Protection Agency framework [\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e].\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec4\" class=\"Section3\"\u003e \u003ch2\u003e1.1.3 Hybrid Models and Advanced AI Techniques\u003c/h2\u003e \u003cp\u003eMachine learning techniques have been widely applied in water quality assessments. An ensemble learning model was used for water quality classification [\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e], while a study in Lam Tsuen River, Hong Kong, utilized the WQI with an Extra Tree regression model [\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e]. Another study employed an XGBoost classifier, achieving 97.06% accuracy using hyperparameter optimization [\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eSeveral hybrid models have been explored to enhance predictive accuracy. A study in the Talar catchment found that a Bagging classifier with a Random Tree outperformed other machine learning models for river quality prediction [\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e]. Comparative studies have evaluated multiple classifiers, including Support Vector Machine (SVM), Random Forest (RF), Logistic Regression (LR), Decision Tree (DT), CATBoost, XGBoost, and Multilayer Perceptron (MLP) [\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e]. The CATBoost model demonstrated the highest accuracy, with feature importance aligning with key pollution indicators [\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e]. Spearman rank correlation coefficients were further used to determine significant trends in pollution indicators [\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eTo facilitate early detection of water contamination, researchers in [\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e] developed a predictive model for BOD₅ values, using linear regression, support vector regression (SVR), and multi-layer perceptrons (MLP) [\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e]. Additionally, various chemometric techniques\u0026mdash;such as Man-Kendall trend analysis, principal component analysis, factor analysis, and agglomerative hierarchical cluster analysis\u0026mdash;have demonstrated the vulnerability of Selangor's water system to ammonia pollutants, posing significant risks to water supply [\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e].\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec5\" class=\"Section3\"\u003e \u003ch2\u003e1.1.4 The Impact of External Factors on Water Quality\u003c/h2\u003e \u003cp\u003eThe impact of COVID-19 on the Ganges River (India) has been studied extensively [\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e]. Various WQI methodologies and water quality parameters have been examined in detail [\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e]. In [\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e]-[\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e], Cascaded Fuzzy Systems were implemented for water quality prediction, while [20] utilized KNN imputation with 10 water quality parameters.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec6\" class=\"Section3\"\u003e \u003ch2\u003e1.1.5 Explainable AI in Water Quality Assessment\u003c/h2\u003e \u003cp\u003eThe integration of Explainable Artificial Intelligence (XAI) in water quality assessment has gained attention in recent years. SHAP (SHapley Additive exPlanations) has been employed to interpret machine learning models, providing transparency in water quality prediction [\u003cspan additionalcitationids=\"CR22 CR23 CR24\" citationid=\"CR20\" class=\"CitationRef\"\u003e21\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e25\u003c/span\u003e]. The use of XAI ensures that AI-driven decisions are interpretable and actionable for stakeholders, improving trust in automated water quality assessments. With traditional Water Quality Index (WQI) value calculation, many errors are introduced; to overcome such errors, we can use machine learning and deep learning models.\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv id=\"Sec7\" class=\"Section2\"\u003e \u003ch2\u003e1.2 Research Gap\u003c/h2\u003e \u003cp\u003eDespite significant advancements in water quality assessment and prediction models, several key gaps remain in the existing literature:\u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003e1. Lack of Explainability in AI-Based Water Quality Predictions\u003c/h3\u003e\n\u003cp\u003eWhile machine learning models such as XGBoost, Random Forest, and neural networks have shown high accuracy in water quality prediction [\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e, \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e, \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e], they often operate as black-box models, making it difficult to understand how predictions are made. Few studies have incorporated Explainable AI (XAI) techniques like SHAP (SHapley Additive ExPlanations) to provide insights into model decisions [\u003cspan additionalcitationids=\"CR22 CR23 CR24\" citationid=\"CR20\" class=\"CitationRef\"\u003e21\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e25\u003c/span\u003e]. However, existing XAI applications in water quality assessment remain limited and underexplored.\u003c/p\u003e\n\u003ch3\u003e2. Insufficient Hybrid and Ensemble Learning Approaches\u003c/h3\u003e\n\u003cp\u003eAlthough individual models such as XGBoost and Support Vector Machines (SVM) have demonstrated promising results [\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e, \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e, \u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e], hybrid models and ensemble learning techniques remain underutilized. Studies have shown that combining multiple models, such as Bagging classifiers and Random Trees [\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e], can enhance predictive performance, but comprehensive evaluations across different water bodies and geographic locations are lacking.\u003c/p\u003e\n\u003ch3\u003e3. Limited Geographic and Environmental Coverage\u003c/h3\u003e\n\u003cp\u003eMost water quality studies have focused on specific rivers and regions, such as:\u003c/p\u003e \u003cp\u003eLam Tsuen River, Hong Kong [\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e]\u003c/p\u003e \u003cp\u003eTalar Catchment, Iran [\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e]\u003c/p\u003e \u003cp\u003eGanges River, India (COVID-19 impact) [\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e]\u003c/p\u003e \u003cp\u003eSelangor, Malaysia (chemometric analysis) [\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e]\u003c/p\u003e \u003cp\u003eHowever, comprehensive datasets covering diverse hydrological conditions and climatic variations are scarce. There is a need for models that generalize across multiple regions and water sources, including groundwater, lakes, and reservoirs.\u003c/p\u003e\n\u003ch3\u003e4. Lack of Real-Time and Early Warning Systems\u003c/h3\u003e\n\u003cp\u003eMost studies rely on historical data for water quality prediction [\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e, \u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e, \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e], rather than real-time monitoring and early warning systems. The integration of IoT sensors, remote sensing, and AI-driven alert mechanisms for water contamination detection remains an open research area.\u003c/p\u003e\n\u003ch3\u003e5. Insufficient Studies on AI-Driven Decision Support Systems\u003c/h3\u003e\n\u003cp\u003eWhile various models assess water quality, few studies explore how AI-based predictions can be effectively integrated into policymaking and public health strategies. There is a lack of user-friendly AI-driven decision support systems that can help environmental agencies and policymakers take preventive actions before contamination reaches critical levels.\u003c/p\u003e\n\u003ch3\u003e6. Data Imbalance and Feature Selection Challenges\u003c/h3\u003e\n\u003cp\u003eMany studies face imbalanced datasets, where instances of extreme pollution events are rare [20]. Furthermore, feature selection methodologies are often inconsistent, leading to suboptimal model performance. There is a need for automated feature selection techniques and strategies to handle data imbalance for more robust AI models.\u003c/p\u003e \u003cp\u003e \u003cb\u003eMotivation and Contribution\u003c/b\u003e \u003c/p\u003e \u003cp\u003eThis research presents a novel approach for waterborne disease detection by integrating efficient machine learning models with Explainable AI (XAI). The key contributions of this study include:\u003c/p\u003e \u003cp\u003eImplementation of advanced machine learning techniques (e.g., XGBoost, Random Forest, and ensemble learning) to classify and predict waterborne disease susceptibility based on water quality parameters.Comparison of various machine learning models to identify the most accurate and efficient approach for disease prediction.\u003c/p\u003e \u003cp\u003eUtilization of SHAP (SHapley Additive ExPlanations) to interpret feature importance, ensuring transparency in water quality assessments and disease prediction outcomes.\u003c/p\u003e \u003cp\u003eIdentification of the most influential water quality parameters contributing to contamination and disease risk, aiding in better decision-making.\u003c/p\u003e \u003cp\u003eApplication of data balancing techniques to handle class imbalances in water quality datasets, improving model robustness.Feature selection and engineering to enhance predictive performance by reducing noise and redundancy in data.\u003c/p\u003e \u003cp\u003ePerformance benchmarking against existing models that use traditional Water Quality Index (WQI) and non-explainable ML approaches.\u003c/p\u003e \u003cp\u003eDemonstration of how XAI improves interpretability and decision support compared to black-box models. Proposal of an AI-driven decision support system to assist environmental agencies and policymakers in early disease outbreak detection.\u003c/p\u003e \u003cp\u003e \u003cb\u003eContribution\u003c/b\u003e:\u003c/p\u003e \u003cp\u003eA pilot study was conducted in western parts of Gujarat, Vadodara an analysis of a number of open and closed gutters in the area was surveyed. Along with 250 people were surveyed for notable diseases related to water were identified and analyzed.\u003c/p\u003e \u003cp\u003eAs per the Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e, study conducted in the western region of India in Gujarat shows the count of disposing the waste in open gutter is more as compared other waste water management areas. The open gutters are an open invitation for water contamination.\u003c/p\u003e \u003cp\u003eFigure \u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e., shows the number of open gutters is more in Gujarat's western region, leading to contamination of water and making it more polluted. Also, the sewage lines in most of the areas remain open, and disposal of wastewater from industries and households leads to severe waterborne diseases. Around 250 people were surveyed for various diseases affected by drinking water or using water repositories near their houses.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eThe Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e. shows us the identified signs and symptoms observed from 250 people who were also surveyed for drinking habits, water reservoirs nearby, and Diarrheal diseases, which were reported and are plotted in the graph.\u003c/p\u003e \u003cp\u003e \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\u003eDiseases caused by associated pathogens\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"7\"\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 \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDiseases\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eBacteria\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eHealth Significance\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eVirus\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eHealth Significance\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003eParasites\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c7\"\u003e \u003cp\u003eHealth Significance\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"4\" rowspan=\"5\"\u003e \u003cp\u003e\u003cb\u003eDiarrhea\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eEscherichia coli (E. coli)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eHigh\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eRotavirus\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eHigh\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eCryptosporidium\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eHigh\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eCampylobacter jejuni\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eHigh\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eNorovirus\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eHigh\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eGiardia\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eHigh\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eSalmonella\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eHigh\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eAdenovirus\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eModerate\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eEntamoeba\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eHigh\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eShigella\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eHigh\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eAstrovirus\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eModerate\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e--\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e--\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eYersinia\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eModerate\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e--\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e--\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eCholera\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eVibrio cholerae\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eHigh\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e--\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e--\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e--\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eHepatitis A\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003ehepatitis A virus (HAV)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eHigh\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e--\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e--\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eTyphoid\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eSalmonella Typhi bacteria\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eHigh\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e--\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e--\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eGiardiasis\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eGiardia duodenalis. Giardia intestinalis.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eHigh\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003ePnemonia\u003c/b\u003e,\u003c/p\u003e \u003cp\u003e\u003cb\u003eMeningitis\u003c/b\u003e,\u003c/p\u003e \u003cp\u003e\u003cb\u003eUrinary Tract infection\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eEnterohemorrhagic E. coli (EHEC)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eHigh\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e--\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e--\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e--\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e---\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eCampylobacteriosis\u003c/b\u003e\u003c/p\u003e \u003cp\u003e\u003cb\u003eGuillain-Barr\u0026eacute; Syndrome\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u0026bull; Campylobacter jejuni.\u003c/p\u003e \u003cp\u003e\u0026bull; Campylobacter coli\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eHigh\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e--\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e--\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e--\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e---\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eCryptosporidiosis\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e---\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e---\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e--\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e--\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eCryptosporidium\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eHigh\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eCyclosporiasis\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e---\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e----\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e--\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e--\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eCyclospora cayetanensis\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eHigh\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eLegionellosis\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eLegionella\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eHigh\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e--\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e--\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e----\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e--\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eShigellosis\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eShigella bacteria\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eHigh\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e--\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e--\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e----\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e--\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eVibriosis\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eVibrio parahaemolyticus and Vibrio vulnificus.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eHigh\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e--\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e--\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e----\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e--\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\u003eTable. 1, describes the diseases associated with the relevant pathogen categorized as Bacteria, Virus and Parasites. Health significance column, relates to the severity of impact with low, moderate and high values, including association with outbreaks.\u003c/p\u003e"},{"header":"2. Materials and Methods","content":"\u003cp\u003eThe dataset is collected from Pollution Board Control consists of more than 2700 data, for 5 years ranging from 2017 to 2022. Each attribute in the dataset consists of the water parameters like pH, BOD, Dissolved solids, Temperature, Conductivity, Nitrogen, Fecal Coliform, and Fecal Streptococci. The dataset consists of river water from various states of India. Overall, 22 states of India were taken into account. All this data was given to the machine learning model for the training phase of the model.\u003c/p\u003e \u003cp\u003eThe data is pre-processed to be fed into the learning model/ classifiers. This data is cleaned with missing values using mean, median, and mode methodology. The data is used for training and testing used in the model. Data studied and utilized from the pilot study also contains the values that were not detected, those values were also removed and can used for testing purposes in the learning model.\u003c/p\u003e \u003cdiv id=\"Sec15\" class=\"Section2\"\u003e \u003ch2\u003e2.1 The Machine learning models\u003c/h2\u003e \u003cp\u003eHistGradientBoosting Classifier- for Big Data sets of more than 10,000 samples, is much faster as compared to GradientBoostingClassifier. This classifier supports the presence of missing values in the dataset. Consider x and y with two inputs that have samples N. The function f(xi), maps the feature x, which is input to the variable y. The summation is given by the following equation The function is called loss function is given by the difference between the actual and the predicted variables.\u003c/p\u003e \u003cp\u003eL(f)=\u0026sum;_(i\u0026thinsp;=\u0026thinsp;1)^N [L(yi,f(xi)] (1)\u003c/p\u003e \u003cp\u003eRandom Forest Classifier - Random Forest supervised machine learning algorithm that combines multiple decision trees to form a forest. Here, GE is the generalization error for the random forest and is denoted as Here, function f (X, Y) is used to count the average of counts and gives predicted value.\u003c/p\u003e \u003cp\u003eGE\u0026thinsp;=\u0026thinsp;Px,y \u0026not;(f (X, Y)\u0026thinsp;\u0026lt;\u0026thinsp;0 (2)\u003c/p\u003e \u003cp\u003eWhere X is called the predicted value, and Y is called the outcome of the classification problem.\u003c/p\u003e \u003cp\u003eBayesian Classifier - Naive Bayes classifier is a machine learning model based on Bayes' theorem [\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e]. It calculates the probability of a given input belonging to a particular class. Here in (3), we have used probabilistic functions to create a classifier model. The probability of given feature inputs for all possible values of the features of y and maximum probability is given as output from the function defined below.\u003c/p\u003e \u003cp\u003ey\u0026thinsp;=\u0026thinsp;max(c)Π_(ⅈ=1)^nc(xi∣y) (3)\u003c/p\u003e \u003cp\u003ec is called the probability of a particular class, and c(xi∣y) is called conditional probability.\u003c/p\u003e \u003cp\u003eDecision tree Classifier- It is a supervised learning technique, hence used for classification and regression for various applications. This Tree classifier has nodes that represent the features of the dataset. Branches indicate the decision rules, and leaves are the outputs of the algorithm.\u003c/p\u003e \u003cp\u003eAdaboosting Classifier- This classifier is used to remove the faults that occur in training the model. It is a machine-learning model used for classification and regression problems.\u003c/p\u003e \u003cp\u003eLong short-term memory (LSTM) model is used. The LSTM is a deep learning model that retains information for a long series time.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec16\" class=\"Section2\"\u003e \u003ch2\u003e2.2 Algorithm\u003c/h2\u003e \u003cp\u003eData preprocessing was done, each of the water quality parameters mentioned had two columns, minimum and maximum.\u003c/p\u003e \u003cp\u003e \u003cul\u003e \u003cli\u003e \u003cp\u003eEach feature of water was segregated with minimum values and trained only for those values. Similarly, every feature of maximum value was trained with ML models.\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003eEach row was initialized with 0/1 catering to susceptible to diseases.\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003eThe dataset comprised of missing values, with the SMOTE method, we evaluated the dataset carefully and adjusted the missing value.\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003eEach trained model was fed with the processed dataset, and each model gave a fairly good performance.\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003eFor the Random Forest classifier, Explainable AI was used since this model gave the highest accuracy using the SHapley Additive exPlanations (SHAP) methodology. SHAP is an analytical approach that explains the output of machine learning model.\u003c/p\u003e \u003c/li\u003e \u003c/ul\u003e \u003c/p\u003e \u003cp\u003eThe figure, shows the entire methodology used for models. All the above steps are also depicted in the figure below.\u003c/p\u003e \u003c/div\u003e"},{"header":"3. Results","content":"\u003cp\u003eFrom the observation of experiments conducted, the classifier described in the methodology section is at par with the dataset consisting of minimum and maximum values. The machine learning algorithms are trained for both minimum and maximum values of water data parameters. Along with the machine learning models, a deep learning model is also trained, which is the long-short-term memory (LSTM) model. The number of layers used is 3, with a dropout value of 0.02. The model used is a sequential model for LSTM. For the LSTM model, the number of epochs used was 15, and we were able to achieve an accuracy of 91.9%.\u003c/p\u003e \u003cp\u003eVarious classifiers are also deployed for the same data set, which gives us the following observed values. Figure\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e, gives us the brief of all the classifiers used for the model, both values of datasets are considered, minimum values and maximum values. And on both values, all the parameters are trained. The following table shows the trained value accuracy for each of the classifiers used.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eFrom the above data, we are also able to calculate the training loss that occurred during the training of the deep learning model. As we can see, deep learning models need huge networked layers, which increases the complexity of the framework.\u003c/p\u003e \u003cp\u003eAlone with machine learning model classifiers, we are able to achieve a sufficient amount of accuracy needed to understand the water data. Each of the tests performed is done on random data picked up from the dataset collected from pilot study.\u003c/p\u003e \u003cdiv id=\"Sec18\" class=\"Section2\"\u003e \u003ch2\u003e3.1. Performance of the models\u003c/h2\u003e \u003cp\u003eHistGradientBoosting Classifier is trained on both values of the dataset, minimum range values of water as well as maximum values. Providing a fair accuracy of 97 to 98%.\u003c/p\u003e \u003cp\u003eRandom Forest outperforms all the other machine learning models. Even though it is one of the traditional methods of evaluating historical data, it gives a good accuracy over the dataset used.\u003c/p\u003e \u003cp\u003eAdaboost, Bagging classifier, and Decision Tree provide fairly good performance in terms of maximum values.\u003c/p\u003e \u003cp\u003eSince the data is also trained for deep learning models like LSTM to provide a fair comparison with machine learning models, LSTM shows an accuracy of around 91.9%, as most of the deep learning models are used for more complex architectures with many hidden layers. This dataset has a varied range right, from minimum values to maximum values for each of the features.\u003c/p\u003e \u003cp\u003eExplainable AI is also called the interpretability of machine learning models used. Machine learning models are thought of as a black box \u003csup\u003e[24\u0026ndash;25]\u003c/sup\u003e. Since the outcome of each model is difficult to trace back to how the results are achieved. Explainable AI helps with the interpretation of each model used. Thus, giving insight into each of the features used and trusting our Machine learning model, which can be safely used over a different range of applications. Since Random Forest is used with XAI, which has outperformed other models, we look forward to getting sustainable results over any other datasets, and we should trust the model in terms of the results obtained.\u003c/p\u003e \u003cp\u003eHere, for Explainable AI, we have implemented and installed SHAP (Shapley Additive exPlanations) with the Python framework. Explainable AI is implemented with SHAP for Random Forest Classifier, which provides an accuracy of 98.53%. Feature evaluation is done with SHAP, which provides very good insights about the model trained and various features applied.\u003c/p\u003e \u003cp\u003eFigure \u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003e., is about the beeswarm model of SHAP, showing high and low values, indicating blue color for low-value impact on the model and red shows high-value impact on the model. Here, we can see temperature shows the high impact on the performance of the Random Forest classifier.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eIn Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003e., the waterfall diagram, the x-axis highlights the values of the dependent variable, which is susceptible to diseases. x is the observed value, f(x)gives the prediction value of the model for a given input x, and E(x) is the expected value of the dependent variable. The average of all predictions is given by (mean(model(f(x))). We can observe for a certain data value; the Total coliform feature is found to be +\u0026thinsp;0.58, having more impact as compared to other features in the dataset.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eFigure \u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003e., gives us a detailed picture of the features making an impact on the overall model. Feature importance can be observed from the plot moving toward the base value to the higher or lower side. As observed, the Total coliform has a much higher impact and is larger plot towards the higher side. For lower predictions, BOD and Nitrate and Nitrite values are observed since the impact on the model is on the lower side. The red color represents high, whereas blue shows low values on correlated variables. In this case, the high positive impact of Total Coliform levels shows susceptibility to diseases from the SHAP values lying on the X-axis.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab2\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eComparative table for Existing models and proposed framework\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"6\"\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 \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eReference\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eWaterbody\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eLocation\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eModels used\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eParameters used\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003eAccuracy\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eD. de Andrade Costa,et.al \u003csup\u003e[1]\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eMixed\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e--\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e(RF) classifier, decision tree (DT) classifier, (AdaBoost) classifier, (SVM), Na\u0026iuml;ve Bayes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e21\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eXGBoost 97.06%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eBhateria, R., Jain, D \u003csup\u003e[13]\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eMixed Water bodies\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eIndia\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eSVM, RF, LR, DT, CATBoost, XGBoost, and MLP [\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e].\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e7\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eCATBoost\u003c/p\u003e \u003cp\u003e(94.51%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMuduli, P.R., Kumar, A., Kanuri, V.V \u003csup\u003e[15]\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eWater reservoirs\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eTaiwan\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eANN\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e6\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e--\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eKambala Vijaya Kumar \u003csup\u003e[19]\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eWater Reservoir\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eGeorgia, USA\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eNeuro-Fuzzy (ANFIS)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e11\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e92.37%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAzween Abdullah,et.al \u003csup\u003e[20]\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eKaggle\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e--\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eDecision Trees, Random Forest, Logistic Regression, Support Vector Machines [20]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e21\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e95.08%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eJinal Patel,et.al, \u003csup\u003e[21]\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eMixed Water\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e--\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eRF, SVM, Gradient Boost\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e9\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e81%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMadni HA, et al., \u003csup\u003e[22]\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eWater\u003c/p\u003e \u003cp\u003e(Kaggle)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e--\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eKNN imputer\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e10\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e97%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eN. Hellen and G. Marvin \u003csup\u003e[23]\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eMixed water bodies\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e---\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eRF,ET,KNN,LDA,SVM\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e10\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e67%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eProposed Work\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u003cb\u003eRiver water\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e\u003cb\u003eIndia\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e\u003cb\u003eHistGradientBoosting Classifier, RF Forest, AdaBoost Classifier, Bagging Classifier, Decision Tree, LSTM\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e\u003cb\u003e9\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u003cb\u003e98.53%\u003c/b\u003e\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\u003eTable\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e shows the comparative study of the existing frameworks with the proposed framework. As we can see in the table, we can achieve good accuracy as compared to existing Machine learning models. One more point, we have also included minimum and maximum values of each of the parameters, which was missing in all the above-studied papers.\u003c/p\u003e \u003c/div\u003e"},{"header":"4. Conclusions","content":"\u003cp\u003eThis study underscores the importance of advanced analytical techniques in water quality assessment and pathogen detection. Through machine learning models, particularly Random Forest and Explainable AI, we successfully identified key contaminants and their impact on public health. The findings emphasize the urgent need for robust water monitoring systems to prevent disease outbreaks and improve water management practices. Future research can extend this work by incorporating microbiological analysis and real-time data integration for predictive water safety measures. By leveraging AI-driven solutions, we can ensure safer water sources and a healthier population.\u003c/p\u003e \u003cp\u003eTo gain insight into the model trained, we were able to use Explainable AI with the Machine learning model to learn about the features that dominate the parameters leading to waterborne diseases.\u003c/p\u003e \u003cp\u003eFurther, we can extend the scope by evaluating the microorganisms found in the water. These micro-organisms could be further compared with the signs and symptoms of people identified with the related diseases observed in the study already carried out.\u003c/p\u003e \u003cp\u003eTo gain insight into the model trained, we were able to use Explainable AI with the Machine learning model to learn about the features that dominate the parameters leading to waterborne diseases.\u003c/p\u003e \u003cp\u003eFurther, we can extend the scope by evaluating the microorganisms found in the water. These micro-organisms could be further compared with the signs and symptoms of people identified with the related diseases observed in the study already carried out.\u003c/p\u003e"},{"header":"Declarations","content":" \u003ch2\u003eConflict of Interest\u003c/h2\u003e \u003cp\u003eAuthors declare no conflict of interest.\u003c/p\u003e \u003ch2\u003eFunding\u003c/h2\u003e \u003cp\u003eAuthors state no funding involved.\u003c/p\u003e\u003ch2\u003eAuthor Contributions\u003c/h2\u003e \u003cp\u003eAuthor -Jaya Zalte, have done cleaning of data, conceptualization, investigation and implementation. Corresponding author- Dr. Harshal shah, have contributed in conceptualization, methodology, inspection and support. Dr. M.H Fulekar have support for the pilot study in Gujarat, he has conceptualised the idea and provided the entire support needed for this project.\u003c/p\u003e\u003ch2\u003eAcknowledgments\u003c/h2\u003e \u003cp\u003eI would like to thank, Pollution Board Control of India and Parul university for providing the support.\u003c/p\u003e\u003ch2\u003eData Availability\u003c/h2\u003e \u003cp\u003eThe data that support the findings of this study are available on request from the corresponding author, [HS]. The data, which contain information that could compromise the privacy of research participants, are not publicly available due to certain restrictions.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003ede Andrade Costa D, Soares de Azevedo JP, Dos Santos MA, Dos Santos Facchetti Vinhaes Assump\u0026ccedil;\u0026atilde;o R (2020) Water quality assessment based on multivariate statistics and water quality index of a strategic river in the Brazilian Atlantic Forest. Sci Rep 10(1):22038. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.nature.com/articles/s41598-020-78563-0\u003c/span\u003e\u003cspan address=\"https://www.nature.com/articles/s41598-020-78563-0\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRam\u0026iacute;rez-Morales D, P\u0026eacute;rez-Villanueva ME, Chin-Pampillo JS, Aguilar-Mora P, Arias-Mora V, Mas\u0026iacute;s-Mora M (2021) Pesticide occurrence and water quality assessment from an agriculturally influenced Latin-American tropical region. Chemosphere. 262,127851 \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.chemosphere.2020.127851\u003c/span\u003e\u003cspan address=\"10.1016/j.chemosphere.2020.127851\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAsadollah SBHS, Sharafati A, Motta D, Yaseen ZM (2021) River water quality index prediction and uncertainty analysis: A comparative study of machine learning models. J Environ Chem Eng 9(1) Article 104599. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.jece.2020.104599\u003c/span\u003e\u003cspan address=\"10.1016/j.jece.2020.104599\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eElvin, Wibowo A (2024) Forecasting water quality through machine learning and hyperparameter optimization. JE Indonesian J Electr Eng Comput Sci 496\u0026ndash;506. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.11591/ijeecs.v33.i1.pp496-506\u003c/span\u003e\u003cspan address=\"10.11591/ijeecs.v33.i1.pp496-506\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBui DT, Khosravi K, Tiefenbacher J, Nguyen H, Kazakis N Improving prediction of water quality indices using novel hybrid machine-learning algorithms. Sci Total Environ, 721:137612, \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.scitotenv.2020.137612,2020\u003c/span\u003e\u003cspan address=\"10.1016/j.scitotenv.2020.137612,2020\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTyagi S et al (2018) Water Quality Assessment in Terms of Water Quality Index. Am J Water Resour 13:4\u0026ndash;38. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.11648/j.ajese.20170104.13\u003c/span\u003e\u003cspan address=\"10.11648/j.ajese.20170104.13\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSocheat Chrea Lo\u0026iuml;c, Tudesque R, Chea (2023) Comparative assessment of water quality classification techniques in the largest north-western river of Cambodia (Sangker River-Tonle Sap Basin), \u003cem\u003eEcological Indicators\u003c/em\u003e. 154. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.ecolind.2023.110759\u003c/span\u003e\u003cspan address=\"10.1016/j.ecolind.2023.110759\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBarzegar Y, Gorelova I, Bellini F, D\u0026rsquo;Ascenzo F Drinking Water Quality Assessment Using a Fuzzy Inference System Method: A Case Study of Rome (Italy). Int J Environ Res Public Health. ; 20(15):6522, \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3390/ijerph20156522,2023\u003c/span\u003e\u003cspan address=\"10.3390/ijerph20156522,2023\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAmeen HA Spring water quality assessment using water quality index in villages of Barwari Bala, Duhok, Kurdistan Region, Iraq. Appl Water Sci, 9, \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s13201-019-1080-z,2019\u003c/span\u003e\u003cspan address=\"10.1007/s13201-019-1080-z,2019\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNida Nasir A, Kansal O, Alshaltone F, Barneih M, Sameer A, Shanableh Ahmed Al-Shamma'a, Water quality classification using machine learning algorithms. J Water Process Eng, 48, \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.jwpe.2022.102920,2022\u003c/span\u003e\u003cspan address=\"10.1016/j.jwpe.2022.102920,2022\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYan T, Zhou A, Shen SL (2023) Prediction of long-term water quality using machine learning enhanced by Bayesian optimization. Environ Pollut 318:120870. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.envpol.2022.120870\u003c/span\u003e\u003cspan address=\"10.1016/j.envpol.2022.120870\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJui-Sheng C, Chia-Chun H, Ha-Son H (2018) Determining the quality of water in the reservoir using machine learning. Ecol Inf 44. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.ecoinf.2018.01.005\u003c/span\u003e\u003cspan address=\"10.1016/j.ecoinf.2018.01.005\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBhateria R, Jain D (2016) Water quality assessment of lake water: a review, Sustain. Water Resour Manag 2:161\u0026ndash;173. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s40899-015-0014-7\u003c/span\u003e\u003cspan address=\"10.1007/s40899-015-0014-7\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLoi JX, Chua ASM, Rabuni MF et al (2022) Water quality assessment and pollution threat to safe water supply for three river basins in Malaysia, \u003cem\u003eSci Total Environ\u003c/em\u003e. ;832:155067. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.scitotenv.2022.155067\u003c/span\u003e\u003cspan address=\"10.1016/j.scitotenv.2022.155067\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e, 2022\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMuduli PR, Kumar A, Kanuri VV et al Water quality assessment of the Ganges River during COVID-19 lockdown. Int J Environ Sci Technol. 18, 1645\u0026ndash;1652, \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1007/s13762-021-03245-x,2021\u003c/span\u003e\u003cspan address=\"10.1007/s13762-021-03245-x,2021\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eUddin MG, Nash S, Olbert AI A review of water quality index models and their use for assessing surface water quality. Ecol Ind, 122, \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.ecolind.2020.107218,2021\u003c/span\u003e\u003cspan address=\"10.1016/j.ecolind.2020.107218,2021\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMagesh NS, Krishnakumar S, Chandrasekar (2013) Groundwater quality assessment using WQI and GIS techniques. Arab J Geosci. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1007/s12517-012-0673-8\u003c/span\u003e\u003cspan address=\"10.1007/s12517-012-0673-8\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eShakour S, Chitsazan M, Mirzaee SY (2023) Zonation of groundwater quality in terms of drinkability, using Fuzzy Logic and Schoeller deterministic method for Northern Dezful - Andimeshk Plain, Iran. Discov Water 3:22. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s43832-023-00046-w\u003c/span\u003e\u003cspan address=\"10.1007/s43832-023-00046-w\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKambala V, Kumar Y, Dileep Kumar SR, Godla MSA, Ansari YAB, El-Ebiary, Muniyandy E (2023) Enhancing Water Quality Forecasting Reliability Through Optimal Parameterization of Neuro-Fuzzy Models via Tunicate Swarm Optimization,\u003cem\u003eInternational Journal of Advanced Computer Science and Applications\u003c/em\u003e, 15(3), DOI: 10.14569/IJACSA.2024.01503110,2024. \u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAzween Abdullah, Himakshi Chaturvedi, Siddhesh Fuladi, Nandhika Jhansi Ravuri, Deepa Natesan and M. K Nallakaruppan, Reliable and Efficient Model for Water Quality Prediction and Forecasting, \u003cem\u003eInternational Journal of Advanced Computer Science and Applications\u003c/em\u003e, 14(12), \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.14569/IJACSA.2023.0141219\u003c/span\u003e\u003cspan address=\"10.14569/IJACSA.2023.0141219\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePatel J, Amipara C, Ahanger TA et al (2022) A Machine Learning-Based Water Potability Prediction Model by Using Synthetic Minority Oversampling Technique and Explainable AI, \u003cem\u003eComput Intell Neurosci\u003c/em\u003e, :9283293. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1155/2022/9283293\u003c/span\u003e\u003cspan address=\"10.1155/2022/9283293\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e, 2022\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMadni HA, Umer M, Ishaq A, Abuzinadah N, Saidani O, Alsubai S, Hamdi M, Ashraf I (2023) Water-Quality Prediction Based on H\u003csub\u003e2\u003c/sub\u003eO AutoML and Explainable AI Techniques. 15(3):475. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3390/w15030475\u003c/span\u003e\u003cspan address=\"10.3390/w15030475\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNakayiza H, Ggaliwango M (2022) Explainable AI for Safe Water Evaluation for Public Health in Urban Settings, International Conference on Innovations in Science, Engineering and Technology (ICISET), 1\u0026ndash;6. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1109/iciset54810.2022.9775912\u003c/span\u003e\u003cspan address=\"10.1109/iciset54810.2022.9775912\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMallick J, Alqadhi S, Hang HT et al (2024) Interpreting optimised data-driven solution with explainable artificial intelligence (XAI) for water quality assessment for better decision-making in pollution management. Environ Sci Pollut Res 31:42948\u0026ndash;42969. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s11356-024-33921-7\u003c/span\u003e\u003cspan address=\"10.1007/s11356-024-33921-7\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRandika K, Makumbura L, Mampitiya N, Rathnayake DPP, Meddage S, Henna TL, Dang Y, Hoshino U, Rathnayake Advancing water quality assessment and prediction using machine learning models, coupled with explainable artificial intelligence (XAI) techniques like Shapley additive explanations (SHAP) for interpreting the black-box nature. Results Eng, 23,2024,102831, ISSN 2590\u0026thinsp;\u0026ndash;\u0026thinsp;1230, \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.rineng.2024.102831\u003c/span\u003e\u003cspan address=\"10.1016/j.rineng.2024.102831\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\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":"Machine Learning, Water Quality, SHAP, Explainable AI Contaminants","lastPublishedDoi":"10.21203/rs.3.rs-6338765/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-6338765/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eWater quality is a critical factor for human health and environmental sustainability. Rapid urbanization and industrialization have led to significant water contamination, increasing the prevalence of waterborne diseases. This study investigates the presence of pathogens in water sources across the Gujarat region, utilizing machine learning models to analyze contamination patterns. Various classifiers, including HistGradientBoosting, Random Forest, AdaBoost, Bagging, Decision Tree, and LSTM, were employed to predict water quality and identify pathogens. Among these, the Random Forest and Bagging classifiers exhibited the highest accuracy at 98.53%. Furthermore, Explainable AI techniques, specifically SHapley Additive exPlanations (SHAP), were used to interpret the significant features influencing contamination levels. The study highlights the need for proactive water quality monitoring and pathogen detection to prevent disease outbreaks.\u003c/p\u003e","manuscriptTitle":"Towards Safer Water: AI-Driven Predictive Analytics for Disease Detection","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-04-01 11:50:57","doi":"10.21203/rs.3.rs-6338765/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":"a09904e0-13cb-4061-952d-2ec8eef4a611","owner":[],"postedDate":"April 1st, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":46413627,"name":"Environmental Engineering"},{"id":46413628,"name":"Health Economics \u0026 Outcomes Research"},{"id":46413629,"name":"Marine and Freshwater Ecology"}],"tags":[],"updatedAt":"2025-04-01T11:50:57+00:00","versionOfRecord":[],"versionCreatedAt":"2025-04-01 11:50:57","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-6338765","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-6338765","identity":"rs-6338765","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}