Adaptive Multilayer Ensemble based Real Time Stream data Classification : Architectural Perspectives and Performance

Adaptive Multilayer Ensemble based Real Time Stream data Classification : Architectural Perspectives and Performance | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Adaptive Multilayer Ensemble based Real Time Stream data Classification : Architectural Perspectives and Performance Shailaja Jadhav, Dattatraya Kodavade This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-5847010/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 Recently, there has been a lot of study focused on large-scale data analysis. Data typically arrives continually in the form of data streams in numerous applications of today’s data intensive world. The analytics engines designed to deal with streaming data need to be capable of acting on data -in- motion . Since the conventional techniques for data mining and machine learning were created for static datasets, data streams present unique difficulties. They are unable to effectively analyse data that is expanding at a quick pace and are less suitable to take into account the representative qualities of data streams. The authors through this research viz. A-MERIT-C ( A daptive M ultitiered E nsemble for R eal ( I ntegrated with feature engineering) T ime C lassification), have presented an active learning data stream analysis model through Ensemble Learning Framework ,which is able to learn concepts in near real time streaming data environment. The design of the framework is adaptive to address the dynamism associated with real time data through the classifier pool (i.e. it chooses the best performing classifiers). This design of Adaptive Multitiered Ensemble is composed of prequentially evaluated classifiers rather than traditional hold out evaluation. The distinguishing characteristics of A-MERIT-C give considerable improvements in terms of f1 (87.48), recall (77.79), accuracy (79.27) AUC (88.89) for streaming data analytics as compared to that of similar real time airline FLTRADAR 24 models ; on the other hand, it is able to address the limitations of existing algorithms such as concept evolution , feature drift through incremental learning and feedback . adaptive ensemble prequential analysis stream data mining streaming data analytics Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 1. Introduction One of the most important success criteria for modern businesses is the ability to base business decisions on knowledge concealed in stored data . Numerous other spheres of human endeavour exhibit comparable curiosity in investigating novel forms of information. Data typically arrives continually in the form of data streams in many of these applications. Examples that exemplify this , include click log mining, analysis of emotions and sentiments, network analysis, financial data prediction, traffic control, sensor measurement processing, ubiquitous computing, GPS and mobile device monitoring and numerous other applications [22]. The value of data has led to the widespread adoption of data technologies today. While the challenge of handling hundreds of gigabytes of data is in place , it is to be taken into account that data generated several months ago has far less value as compared to that generated within an hour. It implies that more recent the data , of higher value it is. Because they are typically generated in a few seconds, or even less than a millisecond, the real-time data are therefore the most important. The major goal of data stream learning is to aptly use limited resources , and create and maintain an up-to-date model which is ingesting speedy data arriving from a stream data source. As stated in Shan et al. [5], Zhai et al. [7], a variety of learning frameworks and methods—mostly ensemble-based—are appropriate choices to deal with data streams and get even greater accuracy. There is an extensive need to design and develop an Ensemble Learning Framework which will learn concepts in real time streaming data environments and will address the limitations of existing algorithms such as concept evolution ,feature drift ,along with improvements in the performance of streaming data analytics. The aim of the research is to design an adaptive classifier ensemble. The remainder of the paper is arranged as follows... A quick know-how of the related work in the similar research area is taken is section 2. Section 3 discusses the main idea behind architecture – design principles , overall system design and the major components. Section 4 is dedicated to the detailed algorithms and pseudocode associated with the major design components. Section 5 illustrates the experimental aspects associated with AMERIT-C , data description and metadata. Following section 6 deals with experimental results and performance analysis as real time stream data classification -flight stream data analysis system. Section 7 summarizes the overall design and experimental observations through concluding remarks. The idea is discussed here through architectural concerns as prime objective and associated part of experimentation ; but the system has major scope for further extension so, section 8 deals with future scope and limitations in the existing research design . 2. Background Since existing approaches were created for static datasets and cannot effectively analyze rapidly increasing amounts of data, data streams present new issues for machine learning and data mining. Supervised classification has drawn the greatest amount of study attention among the various tasks examined in data streams. It is frequently used to address a wide range of real-world issues. The need for novel algorithmic solutions arises from the intrinsic features of stream data, namely its magnitude, speed, and dynamic nature. Because the distribution of the data in motion can change, classifiers specifically designed for data streams must have adaptive capabilities. Streaming data differs from static data in the following ways: • Boundless amount : Since data may be infinite, it is impractical to store all of it for multi-pass processing. Classifiers should operate in situations with restricted resources in order to process this. • High rapidity : Data comes in so fast that a real-time reaction is necessary. In these kinds of settings, algorithms that depend on retraining fresh classifiers on the most recent data could not work well. • Concept drifting : Data has changing distributions rather than being dispersed independently and identically. While the distribution of the observed variables may fluctuate or stay constant over time, the relationship between the observed variables and the goal variable is anticipated to change in an unpredictable manner. Classifiers must be able to forget the past data and remain in tune with the most recent data in order to be impactful on decision-making. Mining changing data streams has garnered a lot of research attention currently in an effort to solve the aforementioned characteristics. Until now, most of the proposed autonomous and supervised data stream mining methods have been developed for classification and clustering tasks. Most of them are based on the standard methods for learning in static data [2,9], with significant modifications to handle data streams. Data stream classification algorithms are based on popular machine learning approaches such as instance-based classifiers, neural networks, Bayesian classifiers, and decision trees [1,6,22]. Because they perform better than depending on strong and stable single learners [8], ensemble-based methods are one of the most popular and often used techniques for data stream classification. Several baseline or primary learners undergo training on the same data in an ensemble, also known as a combination of classifier system, and then their predictions are aggregated. In this manner, the single learners' strengths are merged while their weaknesses are lessened. Furthermore, compared to a single classifier, ensembles are typically more noise-resistant [3, 5]. Enhancing both performance and robustness is the aim of ensemble approaches. The misclassified examples must be diverse for an ensemble to function. Another attribute is its capacity to be readily implemented in practical applications [18]. Ensemble algorithms can be used with drift detection tasks in stream learning and to consider dynamic updates, which include adding classifiers with higher performance and removing classifiers with lower performance [3,7]. For data stream analysis, the ensemble-based approach—a pooled conclusion based on the opinions of a few chosen specialists—is increasingly common. Numerous combination techniques and heuristic approaches [15] are used on a range of fundamental algorithms to address specific issues. There isn't a general guideline for creating the ensemble that yields the optimal solution for each problem [29]. These approaches do not pre-impose the learners to be utilized as component classifiers, examining the most current papers from [8,34] in general. Another effective method for integrating several classifiers is stacked generalization, or stacking. It was first presented by Wolpert and has proven to be highly effective in the classification domain, particularly in cases when the classifiers are only generally accurate in certain regions of the dataset [8]. There are two stages to stacking. Training begins with the base or single learners (level 0). Together with the actual value that those models are meant to forecast, the models' output is gathered into a new dataset [34]. The stacking model learner (level 1), sometimes referred to as the meta-learner, uses this fresh dataset as input before combining it to produce the final output. It is possible to set up a stacking architecture so that the meta-classifier receives the probabilities as input. This approach has the benefit of strengthening inter-level communication because the meta-classifier is aware of the base classifier's confidence [35]. 3. Research Design The kernel-backbone for real time data stream analysis is design of A-MERIT-C ( A daptive M ultitier E nsemble for R eal T ime stream data C lassification - I ntegrated with feature engineering ) . Flight data grabbed with live API from Opensky.org, aviation stack etc. is used for the experiments carried out here. (Section 5.1 ) It undergoes data preprocessing and flight data is made concise enough . Next step is application of feature engineering in order to focus on potential feature set i. e. only the features or characteristics of flight data which contribute to the target data analysis objective.(for details of feature engineering the readers are requested to refer [37] by the authors) The A-MERIT-C is built with the objective of predicting flights if they are delayed or not. Once feature engineering is accomplished, the chunk of data stream is now ready to analyse with A-MERIT-C. The detailed system architecture for the multi-tiered ensemble learning integrated with feature engineering is as shown in fig.1, where every incoming real time data stream undergoes feature engineering and then it becomes a stable input to A-MERIT-C. The major research contribution of this research is dynamic selection of component classifiers to be added in classifier pool. On the fly learning of classifiers takes place here, and our model dynamically selects the best performing classifiers every time a new chunk of data is received. Both, correctness of prediction and time taken to learn are taken into account as the decision parameters for adding a classifier to the classifier pool. (details in sec. 5.2) 3.1 Design principles The main motive behind this research is to tune machine learning algorithms for real time data streams. The design principles followed when designing this system framework can be roughly illustrated as under. 1. In order to overcome the limitations of existing machine learning techniques that are less suitable, if applied for real time data streams ; a more suitable and relevant method prequential evaluation is used. In dynamic learning scenarios, the prequential approach is a majorly applied choice of estimation procedure. The distinguishing merit of this prequential approach comes from training phase incorporated here. There is no need for distinct holdout set, instead of that, all the instances are used in training and testing , thus making maximum use of large and varying nature of incoming data instances , which makes it the most suitable choice for streaming data analysis. 2. Existing machine learning techniques are enhanced with multiple tiers of heterogenous classifiers and then combined with a metaclassifier which in turn is responsible for giving final classification result. 3. Before applying the classifier chain in the form of a multi -tiered ensemble, the real time stream data undergoes a feature engineering approach wherein, data preprocessing takes place and then a set of optimal features is selected in order to reduce data dimensionality that will result in reducing computational complexity of the algorithms applied thereon. 4. The major unique design characteristic of the idea behind the research is getting rid of frequent training of classifiers with every new concept drift of dynamism. The proposed system model of A-MERIT-C tries to overcome this time-consuming task ; by building a classifier pool . The classifier pool is kept ready to predict once it receives an incoming chunk of ready to analyse data. Concept drift adaptation is taken care with inclusion of concept adaptive classifiers such as CVFDT, HAT etc. as one of the classifiers in a classifier pool. 5. One more adaptive way of handling dynamism in real time data is A-MERITC ’s ensemble is not static with fixed set of classifiers. Every time depending on the prequential evaluation of candidate classifiers, only the best performing classifiers are selected to be the part of ensemble. 6. As we are dealing with real time data, the performance decision parameters are, how best the classifier can predict with lesser train time and test time along with the required threshold value of accuracy and precision. 3.2 System Architecture A-MERIT-C The research idea for data stream classification majorly revolves around the core design of adaptive multitiered ensemble which has been designed keeping in mind the prime requirements of dynamism and concept drift handling through best performing classifiers selected on the fly. These component classifiers are kept as a classifier pool for tier 1 and a separate pool for tier 2. The reason behind keeping a separate set of component classifier has a basis from studied literature [10][11] and an objective of speedy analysis at the bottom layer followed by a steady and heterogenous classifier set , at the middle layer. The metaclassifier is chosen with the objective of better handling the earlier predictions and simultaneously doing a speedy analysis. The beauty of all this design again relies on a prequential analysis- i. e. predictive sequential analysis strategy which is most suitable choice when dealing with real time stream data analysis [13][27]. Fig. 2 represents the system architecture and centralised in-depth idea of a A-MERIT-C ( A daptive M ultitier E nsemble for R eal T ime stream data C lassification) . Components of Design The overall design can be divided into sub-modules as given below. The detailed system architecture is given in fig. 2. a. Overall flow of the system - fig 1, Algorithm 4.1 b. Train / build initial A-MERIT-C - with the historical data i.e. where the target labels are present - Right hand side -fig. 2, Algorithm 4.3 – Part - I c. Predict with trained -A-MERIT-C - For new incoming data stream instance - Left Hand side -fig. 2, Algorithm 4.3 – Part – II The proposed strategy aims to train a model on the first stream of data received from the source. After that, for every new stream of data received from the source the model will predict the labels for those examples. Based on the prediction, the performance of the model is investigated and the model is updated. The entire process is repeated for every arrival of a new stream of data from the source. At any instant model dynamically selects the best performing classifiers, which are trained previously through prequential evaluation. A sliding window approach is used since it’s a real time stream data. A concept drift adaptation is handled through continuous learning and feedback mechanism. Classifiers are updated based on the streams of the data received till that instant. Hence the model adapts to the change in the target concept and thus taking care of concept drift or concept evolution in the data stream. As shown in fig. 2, the in-depth architecture of A-MERIT-C can be illustrated as modelling the naïve stage with historical (or real time data at T 0 ) depicted as right-hand side of the figure. and predicting the real time labels using learnt classifiers in building stage , depicted as left-hand side of the figure. Building the model mainly addresses the concept evolution challenge as prequential evaluation approach is focused here, instead of traditional cross validation approach. This approach tries to use hundred percent data for interleaved test-then train (predictive sequential). The model keeps adapting and self-tuning since only the best performing classifiers for a given sliding window of a real time data are always selected and the classifier pool is kept ready. The underlined research uses a multitiered (more than one levels) stacking classifier working on the principle of ensemble learning strategy. There are two tiers with ready set of classifiers waiting for a fresh stream of real time data, built through the building stage of A-MERIT-C. One more distinguishing characteristics is for every next layer, not only the previous predictions are sent but they are also enhanced with the original feature set. As the input is made to stabilise by first passing it through data preprocessing and feature engineering stage, it is also taken care for any feature drift in the data. 4. Design Algorithms This section majorly deals with the major components of the model design through the pseudocode algorithms. 4.1 Notations used in algorithms and figures as applicable: DS: Data Stream HD: Historical Data N, n: Tier 1 classifiers, N: Probable set of candidates, n: selected set of candidates M, m: Tier 2 classifiers, M: Probable set of candidates, m: selected set of candidates TR 1 C: TR 1 C 1 to TR 1 C n -Tier 1 Classifiers TR 1 P: TR 1 P 1 to TR 1 P n -Tier 1 Predictions TR 2 C: TR 2 C 2 to TR 2 C m - Tier 2 Classifiers TR 2 P: TR 2 P 1 toTR 2 P m - Tier 2 Predictions Window: W 1 to Wn (size: 1000 rec.) Metaclassifier: MetaC CC- Component Classifiers - Pool of learned classifiers 4.2 Algorithmic pseudocode Algorithm 4.1 is illustrating overall system functioning sequence and accordingly corresponding sub modules are getting called. Algorithm 4.2 is illustrating building classifier pool initially with historical data and thereafter with recent past sliding window. Part-I is for prequential evaluation and part II is for selecting best performing ones. Algorithm 4.3 - Part I - Update classifiers online-AMERIT-C -Update/Rebuild Algorithm 4.3 - Part II - Predict class labels- AMERIT-C_predict Algorithm 4-1: Overall Working- AMERIT-C Algorithm 4.2- Part I Train AMERIT-C - building classifier pool- I prequential evaluation of classifiers Algorithm 4.2- Part II: selecting best performing classifiers from Part I output, keeping classifier pool ready Train AMERIT-C - building classifier pool- II Algorithm 4.3 Part- I : AMERIT-C -Update/Rebuild 5. Experimental works All the experiments mentioned herein are carried out on the platform scikit multiflow [ 17 ], and Anaconda Jupiter platforms as mentioned in Table 1. Following are the minimum requirements to carry out the experiments. Table 1 Experimental Setup Sr. No. Parameters Values 1 Framework Scikit Multiflow 2 Real Time Stream Data 1. Real time live API from Opensky.org [ 30 ] 2. Real time Live API from Aviation Stack [ 19 ] 3 CPU Intel Core i7 4 GPU A100 from Google Colab , T4 Tesla 16 GDDR6 memory and 2,560 CUDA cores 5 NumPy 1.15.4 6 SciPy 1.1.0 7 Pandas 0.23.4 8 Seaborn and Matplotlib 0.9.0 and 3.0.2, respectively 9 Scikit-Learn 0.20.2 10 Anaconda Navigator Version 5.3.1; x86_64 bit 11 Jupyter Notebook Version 5.7.2; 64 bit 5.1 Data Description and Metadata This research used a real-time flight data collected from opensky.org [ 30 ] It contains data for 583987 flights, with each flight represented by a row in the dataset. The dataset originally had 17 features, which include information such as flight number, origin and destination airport, and flight status. In order to extract more context-sensitive information, we have generated some hierarchical features from the baseline features. Original features are used to create new features that provide additional insights or context. These new features are called extended features. As a result, the dataset used here has 27 features in total, 17 original features and 10 extended features. The extended features are hierarchical features, built on top of the original features. These additional features provide more context-sensitive information and are useful for feature engineering tasks such as feature selection, feature reduction and feature extraction. This data is pre-processed and optimised with different feature engineering algorithms, only the features which are contributing to delay prediction are taken further, rest of the features are not considered. The Live Real time data is grabbed from Open sky and Aviation Stack URL, also the historical data is taken from the link provided via the same API. The historical data is used to train the classifiers and evaluate them prequentially for their possible selection in constructing the heterogeneous ensemble through classifier pool. Every epoch of live data contains about 5k instances of USA geographical area, which is used for testing the system through stacked ensemble classifiers, for prediction of flights delayed or not. In addition to this data, authors of this research have also included following data (Table 2) for experiments in order to check the feasibility of the model and observing models’ performance. Table 2 Data Description taken for experiments 5.2 Experiments We have conducted several experiments with real time data streams for binary classification, to predict flight delays. The scope of this research is limited to only binary classification. Further it can be extended for multiclass classification as well. The major goal of this research is to build the stacked ensemble such that it will find a golden midpoint between the expected accuracy of binary classifiers over a real time continuous data stream and the computing power requirements as well as time & memory requirements simultaneously. This is a very first requirement of any real time stream data analysis system, since the stream data that too if it is live streaming has to undergo data preprocessing, data pipelining and various feature engineering aspects almost on the fly, at least in a single run. Having said this, the system of this type of data analysis should have the prime capacities of giving highest performance consuming as less memory as possible and with very less computing power as far as possible, so that it will not be resource and memory savvy. (Table 3) Table 3 Design of Stacked Ensemble – Selection Criteria Sr. Component Classifiers Parameters/Criteria 1 Tier 1 Hoeffding Tree (HT) Precision, Train_time Extremely Fast Hoeffding Tree (EFDT) Naïve Bayes (NB) Hoeffding Adaptive Tree (HAT) 2 Tier 2 Very Fast Decision Tree (VFDT) Precision, Train_time Very Fast Decision Rules Classifier VFDRC Naïve Bayes (NB) 3 Final Tier (Metaclassifier) GB,LR,ARF Accuracy,AUC-ROC and Quick Response We have evaluated each of the candidate classifiers for their selection into classifier pool of stacked ensemble layer (viz.1,2,3 etc.) and keeping a threshold value for mean precision and train and test time required for each classifier, our model finally selects the classifiers dynamically. Figure 3 (a,b) shows the example execution results of a random epoch execution over a part of the sliding window of real time flight data stream. The similar process is adopted in order to select a metaclassifier, there are directions received from subsequent literature surveys to go for GB-Gradient Boosting, LR- Logistic regression and RF- Random Forest and some variations thereof. We have selected ARF- Adaptive Random Forest since its performance is undoubtedly good when acting as metaclassifier over a big data stream of dynamic nature [ 32 ]. Once the classifiers are trained over a historical data stream, the selected classifiers are added to the classifier pool 1, 2 as per their place and they are used for layer 1,2 predictions. Layer 2 receives layer1 prediction results for reference plus it does its own prediction, these are given as a combined input to layer 3. Where a metaclassifier is giving a result for prediction. Here, it is Adaptive Random Forest selected as a metaclassifier since its higher precision. Once the component classifiers are added to the classifier pool at tier 1 and 2 respectively, we have tested an incoming sliding window of actual real time flight data, (already underwent data preprocessing and feature engineering as shown in the system architecture - refer Fig. 1, and the corresponding data passes through two tiers of ensemble, and then through a metaclassifier before it gives a final classification result for each flight record, whether it belongs to a class of delayed flights (class 1) or to the class of not-delayed flights(class 0). We have conducted the experiments on scikit multiflow and Anaconda Jupiter platforms, (details in table 1) and we have achieved remarkable progress as compared with the similar systems that we found in the literature. Figures 4 and 5 show the representative evaluation we have achieved for A-MERIT-C’s performance over a real time data stream of flights. 6. Results and Performance Analysis We have experimented with real time stream datasets in order to investigate the performance of component classifiers for their selection into ensemble. Table 4 shows performance analysis of component classifiers. As we are dealing with stream data, we have to found out a golden midpoint between classifiers ability to correctly classify an incoming data instance, predicting accurately in very less time as much as possible .So, the evaluation parameters for classifiers performance are kept as accuracy, kappa and time (Table 4). Naïve Bayes classifier although with average accuracy gain, it fits for the requirements of base layer [ 33 ], hoeffding Tree classifier as mentioned in various data stream related research [ 1 ] [ 27 ] is a good choice to be one of the classifiers of an ensemble. Also, KNN variant SAMKNN [ 28] is suitable as it has built in capacity of self-adjusting memory. As we have to design a heterogenous classifiers, the base and middle tier is composed of classifiers which are heterogenous enough, one of the main requirements of ensemble design in order to expect the better performance gain. Table 4 Performance Analysis of component classifiers Component Classifiers of Ensemble Prequential Evaluation Accuracy Kappa Time(Sec) Naïve Bayes Classifier 0.693 0.1513 76.76 SAMKNN - Self-Adjusting Memory KNN 0.8013 0.5902 34.18 Hoeffding Tree 0.7767 0.5394 11.27 Extremely Fast Decision Tree 0.7886 0.5668 467.93 The design of multitiered ensemble relies on a metaclassifier at the final tier, so we need to choose a metaclassifier with optimized performance gain, studies from literature through the works of [ 3 , 5 , 8 , 10 ] have suggested the probable metaclassifiers as Logistic regression [ 21 ], SVM [ 23 ]and ARF[ 36 ]. We have continued our experimentation for a good choice of a metaclassifier. All these results are obtained at intermediate level of ensemble design, but they have played a vital role and laid a strong foundation in the final multitier ensemble design. In order to evaluate A-MERIT-S’s performance, we as researchers have reviewed a long list of relevant research articles but very few of them have really taken into consideration live real time data. Also, it was irrelevant if at all we could have compared the flight data analysis with that of a generic data stream analysis system. In keeping the same objective in mind, we found a similar research work in, where Souza et.al.[ 21 ] have worked with live real time flight data through various logistic regression models. After doing comparative performance analysis with that of [ 21 ], it is clear that our model A-MERIT-C has achieved considerable gain in evaluation parameters like f1, precision, recall. Also, accuracy of our model is at par with that of [ 21 ]. Table 5 shows the example results of this experimentation Table 5 A-MERIT-C performance evaluation Sr. Metric FLRDAR-24 A-MERIT-C 1 F1 72.85 87.48 2 Precision 71.28 100 3 Accuracy 79.57 79.27 4 Recall 75.25 77.75 5 AUC 87.43 88.889 Figure 4 shows the comparative performance analysis where it is clear that model A-MERIT-C has achieved considerable gain in evaluation parameters like f1, precision, recall. Also, Accuracy of the model is at par with that of [ 21 ].Aviation stream data analysis is also taken further with other aviation datasets – on time performance data –as described below. A-MERIT-C with prequential evaluation is tested for the data. It has shown the remarkable performance (see Fig. 5 ) when compared with the benchmarked online active learning ensembles as shown in the table 6. Table 6 Performance evaluation of – AMERIT-C Data Methodology Mean Acc.(%) Mean Time (S) airline-on-time performance data Online AUE 77.06 7.26 DWM 77.29 7.27 ACE 79.61 13.00 A-MERIT-C 80.15 12.50 BOT-US-DATA Online AUE 99.23 30.86 DWM 95.92 23.46 ACE 97.42 45.71 A-MERIT-C 98.11 52.20 It is more obvious from the observations that all stream processing algorithms that too if they are online active learning ensembles, time taken for learning matters most, if accuracy is more. For airline -on time performance data its huge voluminous data approx. 120 million records with approx. 25–30 features, the inherent veracity and volume has limited the accuracy evaluation. On the other hand, Bot-US-Data is balanced enough, so it has taken accuracy to highest level but only at the cost of mean time increased on an average. This observation is very representative as working with real time data, since one has to find out a golden midpoint between time and accuracy. So, the models should be able to predict in as less time as possible and will be able to produce commendable progress in accuracy. It is also possible to evaluate the system for other performance parameters such as validation loss, MSE, prequential AUC, but the scope for the paper is kept limited only to the accuracy, precision and mean time taken approx. The research work is still continued for the extension. 7. Concluding remarks and discussions This research is very primitive attempt to address near real time stream data classification task w.r.t. aviation streams. There are some observations which will be definitely helpful for the novice researchers at least , who wish to work in the similar area. 1. In case of live real time data, feature engineering and feature optimization play a very vital role in order to prepare data for analysis and to reduce dimensionality , which in turn further leads to minimum computational overhead for data analysis algorithms. 2. Using Stacking with three layers serves better as long as time required for analysis and computational efficiency is considered. 3. Concept Drift Adaptation is widely taken care of and contributes in great extent when dealing with dynamic real time data stream. This is made possible via incrementally learning the algorithms in classifier pool through recent past historical data and always selecting the best performing classifiers to be the part of stacked ensemble learning. 4. Also, metaclassifiers like ARF used here are capable of dealing with drift adaptation. It is an added advantage of A-MERIT-C. 5. Keeping a learnt classifier always ready in a classifier pool leads to reducing the time required to train the classifiers drastically, as decision tress and random forests although yield very good performance , they have high demand for train time. In order to tackle with this issue , A-MERIT-C always keeps the pool of classifiers , ready to test an incoming data stream on the fly. 8. Constraints, limitations and Future scope It’s a very true observation as working with the real time data stream analysis that , there is still scarcity of benchmarked platforms for computation. Although, real time stream analysis is need of the hour , not a plenty of standardised algorithms are being explored by research community ; (with very few exceptions). So, the desire of building a real time system concludes with actually building near-real time when facing challenges in data stream analysis[22,24] . As all real time systems are to be greatly improved with human intervention , same is the case with this research, if added with human intervention-intelligence , this can be of great help. This research lacks this aspect and can be taken further. One more aspect to be noted here is , this research does not take into consideration any weather delays, or catastrophic incidents in aviation stream analysis just to keep the system complexity to the minimum level. This can be taken for further extension of the research. Further work can be inclusion of parallelization of some classifiers ---in GPU and other scalable computing environments (Scalable ensemble). Declarations Author Contribution Shailaja Jadhav , a corresponding author has conducted research , prepared manuscript .D. Kodavade has reviewed , corrected and guided the manuscript writing . References Souza, V.M.A., dos Reis, D.M., Maletzke, A.G. et al. : Challenges in benchmarking stream learning algorithms with real-world data. Data Mining and Knowledge Discovery 34, 1805–1858 (2020) Leite, D., Škrjanc, I. & Gomide, F. An overview on evolving systems and learning from stream data. Evolving Systems 11, 181–198 (2020). https://doi.org/10.1007/s12530-020-09334-5 Jicheng Shan , Hang Zhang, Weike Liu, and Qingbao Liu- “Online Active Learning Ensemble Framework for Drifted Data Streams” - IEEE Transactions on Neural Networks and Learning systems.vol – 6 Issue 1- 2162-237X © 2018 Qi Wang, ZhiHao Luo, JinCai Huang, YangHe Feng, and Zhong Liu “A Novel Ensemble Method for Imbalanced Data Learning: Bagging of Extrapolation-SMOTE SVM ”– Hindawi Computational Intelligence and Neuroscience Volume 2017 Heitor Murilo Gomes, Jean Paul Barddal, and Fabr´icio enembreck , A Survey on Ensemble Learning for Data Stream Classification”, ACM Computing Surveys, Vol. 50, No. 2, Article 23, Publication date: March 2017 Leszek Rutkowski, Maciej Jaworski, Piotr Duda, “Stream Data Mining: Algorithms and Their Probabilistic Properties ” ISBN 978-3-030-13961-2 eBook Tingting Zhai– “ Classification of high-dimensional evolving data streams via a resource-efficient online ensemble” –Springer - Data Mining and Knowledge Discovery September 2017, Volume 31, Issue 5, pp 1242–1265 Dariusz Brzezinski, Jerzy Stefanowski, “Combining block-based and online methods in learning ensembles from concept drifting data streams”, Information Sciences, Volume 265, 2014, Pages 50-67 Leite, D., Škrjanc, I. & Gomide, F. An overview on evolving systems and learning from stream data. Evolving Systems 11, 181–198 (2020). Heitor Murilo Gomes, Jean Paul Barddal, and Fabr´icio enembreck , “ A Survey on Ensemble Learning for Data Stream Classification”, ACM Computing Surveys, Vol. 50, No. 2, Article 23, Publication date: March 2017 Bartosz Krawczyk, Leandro L. Minku, João Gama, Jerzy Stefanowski, MichałWoźniak, “Online Active Learning Ensemble Framework for Drifted Data Streams”, Information Fusion, Volume 37,2017, Pages 132-156 Minku, Leandro & Yao, Xin. DDD: A New Ensemble Approach for Dealing with Concept Drift. Knowledge and Data Engineering, IEEE Transactions on. 24. 619 – 633, (2012). Brzezinski, D., Stefanowski, J. , Prequential AUC: properties of the area under the ROC curve for data streams with concept drift. Knowledge Information Systems 52, 531–562 (2017). Elwell R. and Polikar R., “Incremental Learning of Concept Drift in Nonstationary Environments“, IEEE Trans. on Neural Networks, vol. 22, 1517-1531, 2011. Bertini Junior, J.R., do Carmo Nicoletti,M.: An iterative boosting-based ensemble for streaming data classification. Inf. Fusion 45, 66–78 (2019) Krawczyk, B., Cano, A.: Online ensemble learning with abstaining classifiers for drifting and noisy data streams. Appl. Soft Comput. 68, 677–692 (2018) Scikit-multiflow :Montiel, J., Read, J., Bifet, A., & Abdessalem, T. (2018). Scikit-multiflow: A multi-output streaming framework. The Journal of Machine Learning Research, 19(72):1−5. Z. Yu et al., "Hybrid Incremental Ensemble Learning for Noisy Real-World Data Classification," in IEEE Transactions on Cybernetics, vol. 49, no. 2, pp. 403-416, Feb. 2019, doi: 10.1109/TCYB.2017.2774266 AviationStack: https://aviationstack.com Kluyver, T. et al., 2016. Jupyter Notebooks – a publishing format for reproducible computational workflows. In F. Loizides & B. Schmidt, eds. Positioning and Power in Academic Publishing: Players, Agents and Agendas. pp. 87–90. Aljubairy, A., Zhang, W.E., Shemshadi, A. et al. A system for effectively predicting flight delays based on IoT data. Computing 102, 2025–2048 (2020). Souza, V.M.A., dos Reis, D.M., Maletzke, A.G. et al. Challenges in benchmarking stream learning algorithms with real-world data. Data Mining and Knowledge Discovery, vol. 34, 1805–1858 (2020). Qi Wang, ZhiHao Luo, JinCai Huang, YangHe Feng, and Zhong Liu “A Novel Ensemble Method for Imbalanced Data Learning: Bagging of Extrapolation-SMOTE SVM ”– Hindawi Computational Intelligence and Neuroscience Volume 2017, Article ID 1827016 Alonso-Betanzos, Amparo ,Pavel, Rokas, Kurasova. Prediction of Flight Time Deviation for Lithuanian Airports Using Supervised Machine Learning Model, Computational Intelligence and Neuroscience, Hindawi, SP-8878681,VL - 2020 Gomes, H.M., Bifet, A., Read, J. et al. Adaptive random forests for evolving data stream classification. Mach Learn 106, 1469–1495 (2017). Kolter, J. Z., & Maloof, M. A. Dynamic Weighted Majority: An ensemble method for drifting concepts. Journal of Machine Learning Research 8:2755–2790. (2007). Gonzalez Hidalgo, Juan & Maciel, Bruno Iran Ferreira & Barros, Roberto. Experimenting with prequential variations for data stream learning evaluation. Computational Intelligence. 35. 670-692, (2019). H.-W. Hu, Y.-L. Chen, K. Tang, A dynamic discretization approach for constructing decision trees with a continuous label, IEEE Trans. Knowl. Data Eng. 21 (11) (2009) 1505–1514. V. Losing, B. Hammer and H. Wersing, "KNN Classifier with Self Adjusting Memory for Heterogeneous Concept Drift," 2016 IEEE 16th International Conference on Data Mining (ICDM), Barcelona, Spain, 2016 D. Brzezinski, J. Stefanowski, R. Susmaga and I. Szczech, "On the Dynamics of Classification Measures for Imbalanced and Streaming Data," in IEEE Transactions on Neural Networks and Learning Systems , vol. 31, no. 8, pp. 2868-2878, Aug. 2020. Matthias Schäfer, Martin Strohmeier, Vincent Lenders, Ivan Martinovic and Matthias Wilhelm. "Bringing Up OpenSky: A Large-scale ADS-B Sensor Network for Research". In Proceedings of the 13th IEEE/ACM International Symposium on Information Processing in Sensor Networks (IPSN), pages 83-94, April 2014. Airline On-time Performance Data : https://www.kaggle.com/datasets/bulter22/airline-data Gomes, H.M., Bifet, A., Read, J. et al. Adaptive random forests for evolving data stream classification. Mach Learn 106, 1469–1495 (2017). Y. Yang, G.I. Webb, Discretization for Naive–Bayes learning: managing discretization bias and variance, Mach. Learn. 74 (1) (2009) 39–74. J. Kunnen, M. Duchateau, Z. Van Veldhoven and J. Vanthienen, Benchmarking Stacking Against Other Heterogeneous Ensembles in Telecom Churn Prediction, 2020 IEEE Symposium Series on Computational Intelligence Liu, N., Gao, H., Zhao, Z. et al. A stacked generalization ensemble model for optimization and prediction of the gas well rate of penetration: a case study in Xinjiang. J Petrol Explor Prod Technol 12 , 1595–1608 (2022). Gomes, H.M., Bifet, A., Read, J. et al. Adaptive random forests for evolving data stream classification. Mach Learn 106, 1469–1495 (2017). Jadhav, Shailaja & Kodavade, D.. (2023). Enhancing Flight Delay Prediction through Feature Engineering in Machine Learning Classifiers: A Real Time Data Streams Case Study. International Journal on Recent and Innovation Trends in Computing and Communication. 11. 212-218. 10.17762/ijritcc.v11i2s.6064. Additional Declarations No competing interests reported. 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-5847010","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":405949811,"identity":"e5497752-9ebd-4cd1-9c4a-d9439259d26f","order_by":0,"name":"Shailaja Jadhav","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA4UlEQVRIiWNgGAWjYBACxgYGMxCdwA8XOgBGRGiRbCBWCxBAtBjAlRFSz9x+eNuDn211ecbXTid++FDBIMd3I4HxcAE+h/WklRv2th0uNrudu1lyxhkGY8kbCQyHZ+DT0pBjJsFz5kDittu525h52xgSN4C08ODT0v/GTPLPmbrEzbNBWv4x1BPWMiPHTJqngjlxgzRISwMwHAhreVYmLVNxuFgC7JdjEoYzzzxswKvFsD95m+Qbg7o8/tm5Gz98qLGR5zuefPgzXi0NqHwJBnD04gPyeGVHwSgYBaNgFIAAACg3Uy+iTgrUAAAAAElFTkSuQmCC","orcid":"","institution":"Shivaji University","correspondingAuthor":true,"prefix":"","firstName":"Shailaja","middleName":"","lastName":"Jadhav","suffix":""},{"id":405949812,"identity":"2ef6e633-e87d-4532-b71d-64863ca70046","order_by":1,"name":"Dattatraya Kodavade","email":"","orcid":"","institution":"Shivaji University","correspondingAuthor":false,"prefix":"","firstName":"Dattatraya","middleName":"","lastName":"Kodavade","suffix":""}],"badges":[],"createdAt":"2025-01-17 07:38:20","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-5847010/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-5847010/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":75902956,"identity":"5fbd0555-ac23-430f-b0f5-f4c27b5b7918","added_by":"auto","created_at":"2025-02-10 11:33:15","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":178066,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eOverall Block diagram\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"1.png","url":"https://assets-eu.researchsquare.com/files/rs-5847010/v1/8f85e007219610b0a3c08e51.png"},{"id":75902957,"identity":"edb1ad0f-4823-4956-a405-04728d157216","added_by":"auto","created_at":"2025-02-10 11:33:15","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":110921,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eArchitecture: Adaptive Multitiered Ensemble for Real Time Classification [A-MER(I)T-C]\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"2.png","url":"https://assets-eu.researchsquare.com/files/rs-5847010/v1/58db34e80b7db175d2c28d93.png"},{"id":75904586,"identity":"b8d3f9ba-b4bc-4211-9c35-f3534b865d54","added_by":"auto","created_at":"2025-02-10 11:41:15","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":69450,"visible":true,"origin":"","legend":"\u003cp\u003e(a) optimum size selection for sliding window (b) Performance evaluation of component classifiers to be in dynamic ensemble\u003c/p\u003e","description":"","filename":"3.png","url":"https://assets-eu.researchsquare.com/files/rs-5847010/v1/85b60b73f4ba24a2c5475a62.png"},{"id":75904584,"identity":"68f4120e-0e3a-42fa-a24f-2c741d972503","added_by":"auto","created_at":"2025-02-10 11:41:15","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":31030,"visible":true,"origin":"","legend":"\u003cp\u003eA-MERIT-C performance evaluation with another flight data Analysis system\u003c/p\u003e","description":"","filename":"4.png","url":"https://assets-eu.researchsquare.com/files/rs-5847010/v1/c6d96c79d12a3b3d2ae9fc90.png"},{"id":75902961,"identity":"edfa9107-ced0-46e1-9081-8821ba3b4d01","added_by":"auto","created_at":"2025-02-10 11:33:15","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":59974,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003ePerformance evaluation of A-MERIT-C (Refer data description from table 2\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"5.png","url":"https://assets-eu.researchsquare.com/files/rs-5847010/v1/ad9dc474e549c3d4583d87f8.png"},{"id":76355376,"identity":"54304876-d77f-4194-bab3-a6457df03562","added_by":"auto","created_at":"2025-02-15 11:46:48","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1573576,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-5847010/v1/992e6d08-f4e7-46f6-8ad8-c1c99762b415.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Adaptive Multilayer Ensemble based Real Time Stream data Classification : Architectural Perspectives and Performance","fulltext":[{"header":"1.\tIntroduction","content":"\u003cp\u003eOne of the most important success criteria for modern businesses is the ability to base business decisions on knowledge concealed in stored data . Numerous other spheres of human endeavour exhibit comparable curiosity in investigating novel forms of information. Data typically arrives continually in the form of data streams in many of these applications. Examples that exemplify this , include click log mining, analysis of emotions and sentiments, network analysis, financial data prediction, traffic control, sensor measurement processing, ubiquitous computing, GPS and mobile device monitoring and numerous other applications [22]. The value of data has led to the widespread adoption of data technologies today. While the challenge of handling hundreds of gigabytes of data is in place , it is to be taken into account \u0026nbsp;that data generated several months ago has far less value as compared to that generated within an hour. It implies that more recent the data , of higher value it is. Because they are typically generated in a few seconds, or even less than a millisecond, the real-time data are therefore the most important.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eThe major goal of data stream learning is to aptly use limited resources , and create and maintain an up-to-date model which is ingesting speedy data \u0026nbsp;arriving from a stream data source. As stated in Shan et al. [5], Zhai et al. [7], a variety of learning frameworks and methods—mostly ensemble-based—are appropriate choices to deal with data streams and get even greater accuracy. There is an extensive need to design and develop an Ensemble Learning Framework which will learn concepts in real time streaming data environments and will address the limitations of existing algorithms such as concept evolution ,feature drift ,along with improvements in the performance of streaming data analytics. The aim of the research is to design an adaptive classifier ensemble.\u003c/p\u003e\n\u003cp\u003eThe remainder of the paper is arranged as follows... A quick know-how of the related work in the similar research area is taken is section 2. \u0026nbsp;Section 3 discusses the main idea behind architecture – design principles , overall system design and the major components. Section 4 is dedicated to the detailed algorithms and pseudocode associated with the major design components. Section 5 illustrates the experimental aspects associated with AMERIT-C , data description and metadata. Following section 6 deals with experimental results and performance analysis as real time stream data classification -flight stream data analysis system. Section 7 summarizes the overall design and experimental observations \u0026nbsp;through concluding remarks. The idea is discussed here through architectural concerns as prime objective and associated part of experimentation ; but the system has major scope for further extension so, section 8 deals with future scope and limitations in the existing research design . \u0026nbsp;\u003c/p\u003e"},{"header":"2.\tBackground ","content":"\u003cp\u003eSince existing approaches were created for static datasets and cannot effectively analyze rapidly increasing amounts of data, data streams present new issues for machine learning and data mining. Supervised classification has drawn the greatest amount of study attention among the various tasks examined in data streams. It is frequently used to address a wide range of real-world issues. The need for novel algorithmic solutions arises from the intrinsic features of stream data, namely its magnitude, speed, and dynamic nature. Because the distribution of the data in motion can change, classifiers specifically designed for data streams must have adaptive capabilities.\u0026nbsp;\u003cbr\u003e\u0026nbsp;Streaming data differs from static data in the following ways:\u003c/p\u003e\n\u003cp\u003e•\u0026nbsp;\u003cstrong\u003eBoundless amount\u003c/strong\u003e: Since data may be infinite, it is impractical to store all of it for multi-pass processing. Classifiers should operate in situations with restricted resources in order to process this.\u0026nbsp;\u003cbr\u003e•\u0026nbsp;\u003cstrong\u003eHigh rapidity\u003c/strong\u003e: Data comes in so fast that a real-time reaction is necessary. In these kinds of settings, algorithms that depend on retraining fresh classifiers on the most recent data could not work well.\u003cbr\u003e• \u003cstrong\u003eConcept drifting\u003c/strong\u003e: Data has changing distributions rather than being dispersed independently and identically. While the distribution of the observed variables may fluctuate or stay constant over time, the relationship between the observed variables and the goal variable is anticipated to change in an unpredictable manner. Classifiers must be able to forget the past data and remain in tune with the most recent data in order to be impactful on decision-making.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eMining changing data streams has garnered a lot of research attention currently in an effort to solve the aforementioned characteristics. Until now, most of the proposed autonomous and supervised data stream mining methods have been developed for classification and clustering tasks. Most of them are based on the standard methods for learning in static data [2,9], with significant modifications to handle data streams. Data stream classification algorithms are based on popular machine learning approaches such as instance-based classifiers, neural networks, Bayesian classifiers, and decision trees [1,6,22].\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eBecause they perform better than depending on strong and stable single learners [8], ensemble-based methods are one of the most popular and often used techniques for data stream classification. Several baseline or primary learners undergo training on the same data in an ensemble, also known as a combination of classifier system, and then their predictions are aggregated. In this manner, the single learners' strengths are merged while their weaknesses are lessened. Furthermore, compared to a single classifier, ensembles are typically more noise-resistant [3, 5]. Enhancing both performance and robustness is the aim of ensemble approaches. The misclassified examples must be diverse for an ensemble to function. Another attribute is its capacity to be readily implemented in practical applications [18].\u003c/p\u003e\n\u003cp\u003eEnsemble algorithms can be used with drift detection tasks in stream learning and to consider dynamic updates, which include adding classifiers with higher performance and removing classifiers with lower performance [3,7]. For data stream analysis, the ensemble-based approach—a pooled conclusion based on the opinions of a few chosen specialists—is increasingly common. Numerous combination techniques and heuristic approaches [15] are used on a range of fundamental algorithms to address specific issues. There isn't a general guideline for creating the ensemble that yields the optimal solution for each problem [29]. These approaches do not pre-impose the learners to be utilized as component classifiers, \u0026nbsp;examining the most current papers from [8,34] in general.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eAnother effective method for integrating several classifiers is stacked generalization, or stacking. It was first presented by Wolpert and has proven to be highly effective in the classification domain, particularly in cases when the classifiers are only generally accurate in certain regions of the dataset [8]. There are two stages to stacking. Training begins with the base or single learners (level 0). Together with the actual value that those models are meant to forecast, the models' output is gathered into a new dataset [34]. The stacking model learner (level 1), sometimes referred to as the meta-learner, uses this fresh dataset as input before combining it to produce the final output. It is possible to set up a stacking architecture so that the meta-classifier receives the probabilities as input. This approach has the benefit of strengthening inter-level communication because the meta-classifier is aware of the base classifier's confidence [35].\u003c/p\u003e"},{"header":"3.\tResearch Design ","content":"\u003cp\u003eThe kernel-backbone for real time data stream analysis is design of\u003cstrong\u003e\u0026nbsp;A-MERIT-C (\u003cu\u003eA\u003c/u\u003e\u003c/strong\u003edaptive \u003cstrong\u003e\u003cu\u003eM\u003c/u\u003e\u003c/strong\u003eultitier \u003cstrong\u003e\u003cu\u003eE\u003c/u\u003e\u003c/strong\u003ensemble for \u003cstrong\u003e\u003cu\u003eR\u003c/u\u003e\u003c/strong\u003eeal \u003cstrong\u003eT\u003c/strong\u003eime stream data \u003cstrong\u003e\u003cu\u003eC\u003c/u\u003e\u003c/strong\u003elassification - \u003cstrong\u003eI\u003c/strong\u003entegrated with feature engineering ) . Flight data grabbed with live API from Opensky.org, aviation stack etc. is used for the experiments carried out here. (Section 5.1 ) It undergoes data preprocessing and flight data is made concise enough . Next step is application of feature engineering in order to focus on potential feature set i. e. only the features or characteristics of flight data which contribute to the target data analysis objective.(for details of feature engineering the readers are requested to refer [37] by the authors) The A-MERIT-C is built with the objective of predicting flights if they are delayed or not. Once feature engineering is accomplished, the chunk of data stream is now ready to analyse with A-MERIT-C. \u0026nbsp; The detailed system architecture for the multi-tiered ensemble learning integrated with feature engineering is as shown in fig.1, where every incoming real time data stream undergoes feature engineering and then it becomes a stable input to A-MERIT-C.\u003c/p\u003e\n\u003cp\u003eThe major research contribution of this research is dynamic selection of component classifiers to be added in classifier pool. On the fly learning of classifiers takes place here, and our model dynamically selects the best performing classifiers every time a new chunk of data is received. Both, correctness of prediction and time taken to learn are taken into account as the decision parameters for adding a classifier to the classifier pool. (details in sec. 5.2)\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e3.1 \u0026nbsp;Design principles\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe main motive behind this research is to tune machine learning algorithms for real time data streams. The design principles followed when designing this system framework can be roughly illustrated as under.\u003c/p\u003e\n\u003cp\u003e1. In order to overcome the limitations of existing machine learning techniques that are less suitable, if applied for real time data streams ; a \u003cstrong\u003e\u003cem\u003emore suitable and relevant method prequential evaluation\u003c/em\u003e\u003c/strong\u003e is used. In dynamic learning scenarios, the prequential approach is a majorly applied choice of \u0026nbsp;estimation procedure. The distinguishing merit of this prequential approach comes from training phase incorporated here. There is no need for \u0026nbsp; distinct \u0026nbsp;holdout set, instead of that, all the instances are used in training and testing , thus making \u003cstrong\u003e\u003cem\u003emaximum use of large and varying nature of incoming data instances\u003c/em\u003e\u003c/strong\u003e, which makes it the most suitable choice for streaming data analysis.\u003c/p\u003e\n\u003cp\u003e2. Existing machine learning techniques are \u003cstrong\u003e\u003cem\u003eenhanced with multiple tiers of heterogenous classifiers\u003c/em\u003e\u003c/strong\u003e and then combined with a metaclassifier which in turn is responsible for giving final classification result.\u003c/p\u003e\n\u003cp\u003e3. Before applying the classifier chain in the form of a multi -tiered ensemble, the real time stream data undergoes a \u003cstrong\u003e\u003cem\u003efeature engineering approach\u003c/em\u003e\u003c/strong\u003e wherein, data preprocessing takes place and then a set of optimal features is selected in order to reduce data dimensionality that will result in reducing computational complexity of the algorithms applied thereon.\u003c/p\u003e\n\u003cp\u003e4. The major unique design characteristic of the idea behind the research is getting \u0026nbsp;rid of frequent training of classifiers with every new concept drift of dynamism. The proposed system model of A-MERIT-C tries to overcome this time-consuming task ; by \u0026nbsp;\u003cstrong\u003e\u003cem\u003ebuilding a classifier pool\u003c/em\u003e\u003c/strong\u003e . The classifier pool \u0026nbsp;is kept ready to predict once it receives an incoming chunk of ready to analyse data. Concept drift adaptation is taken care with inclusion of \u003cstrong\u003e\u003cem\u003econcept adaptive classifiers such as CVFDT, HAT\u003c/em\u003e\u003c/strong\u003e etc. as one of the classifiers in a classifier pool.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e5. One more adaptive way of handling dynamism in real time data is A-MERITC \u0026rsquo;s ensemble is not static with fixed set of classifiers. Every time depending on the prequential evaluation of candidate classifiers, only the best performing classifiers are selected to be the part of ensemble.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e6. As we are dealing with real time data, the performance decision parameters are, how best the classifier can predict with lesser train time and test time along with the required threshold value of accuracy and precision.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e3.2 \u0026nbsp;System Architecture A-MERIT-C\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe research idea for data stream classification majorly revolves around the core design of adaptive multitiered ensemble which has been designed keeping in mind the prime requirements of dynamism and concept drift handling through best performing classifiers selected on the fly. These component classifiers are kept as a classifier pool for tier 1 and a separate pool for tier 2. The reason behind keeping a separate set of component classifier has a basis from studied literature [10][11] and an objective of speedy analysis at the bottom layer followed by a steady and heterogenous classifier set , at the middle layer. The metaclassifier is chosen with the objective of better handling the earlier predictions and simultaneously doing a speedy analysis. The beauty of all this design again relies on a prequential analysis- i. e. \u0026nbsp; predictive sequential analysis strategy which is most suitable choice when dealing with real time stream data analysis [13][27]. Fig. 2 represents the system architecture and centralised in-depth idea of a A-MERIT-C (\u003cstrong\u003e\u003cu\u003eA\u003c/u\u003e\u003c/strong\u003edaptive \u003cstrong\u003e\u003cu\u003eM\u003c/u\u003e\u003c/strong\u003eultitier \u003cstrong\u003e\u003cu\u003eE\u003c/u\u003e\u003c/strong\u003ensemble for \u003cstrong\u003e\u003cu\u003eR\u003c/u\u003e\u003c/strong\u003eeal \u003cstrong\u003eT\u003c/strong\u003eime stream data \u003cstrong\u003e\u003cu\u003eC\u003c/u\u003e\u003c/strong\u003elassification) .\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eComponents of Design\u0026nbsp;\u003c/strong\u003eThe overall design can be divided into sub-modules as given below. The detailed system architecture is given in fig. 2.\u003c/p\u003e\n\u003cp\u003ea. Overall flow of the system\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e- fig 1, Algorithm 4.1\u003c/p\u003e\n\u003cp\u003eb. Train / build initial A-MERIT-C\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e- with the historical data i.e. where the target labels are present\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e- Right hand side -fig. 2, Algorithm 4.3 \u0026ndash; Part - I\u003c/p\u003e\n\u003cp\u003ec. Predict with trained -A-MERIT-C\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e- \u0026nbsp; For new incoming data stream instance \u0026nbsp;\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e- Left Hand side -fig. 2, Algorithm 4.3 \u0026ndash; Part \u0026ndash; II\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eThe proposed strategy aims to train a model on the first stream of data received from the source. After that, for every new stream of data received from the source the model will predict the labels for those examples. Based on the prediction, \u0026nbsp;the performance of the model is investigated and the model is updated. The entire process is repeated for every arrival of a new stream of data from the source. At any instant model dynamically selects the best performing classifiers, which are trained previously through prequential evaluation. A sliding window approach is used since it\u0026rsquo;s a real time stream data. A concept drift adaptation is handled through continuous learning and feedback mechanism. Classifiers are updated based on the streams of the data received till that instant. Hence the model adapts to the change in the target concept and thus taking care of concept drift or concept evolution in the data stream.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eAs shown in fig. 2, the in-depth architecture of A-MERIT-C can be illustrated as modelling the na\u0026iuml;ve stage with historical (or real time data at T\u003csub\u003e0\u0026nbsp;\u003c/sub\u003e) depicted as right-hand side of the figure. and predicting the real time labels using learnt classifiers in building stage , depicted as left-hand side of the figure. Building the model mainly addresses the concept evolution challenge as prequential evaluation approach is focused here, instead of traditional cross validation approach. This approach tries to use hundred percent data for interleaved test-then train (predictive sequential). The model keeps adapting and self-tuning since only the best performing classifiers for a given sliding window of a real time data are always selected and the classifier pool is kept ready. The underlined research uses a multitiered (more than one levels) stacking classifier working on the principle of ensemble learning strategy. There are two tiers with ready set of classifiers waiting for a fresh stream of real time data, built through the building stage of A-MERIT-C. One more distinguishing characteristics is for every next layer, not only the previous predictions are sent but they are also enhanced with the original feature set. As the input is made to stabilise by first passing it through data preprocessing and feature engineering stage, it is also taken care for any feature drift in the data.\u003c/p\u003e"},{"header":"4. Design Algorithms","content":"\u003cp\u003eThis section majorly deals with the major components of the model design through the pseudocode algorithms.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e4.1 Notations used in algorithms and figures as applicable:\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eDS: Data Stream HD: Historical Data\u003c/p\u003e\n\u003cp\u003eN, n: Tier 1 classifiers, N: Probable set of candidates, n: selected set of candidates\u003c/p\u003e\n\u003cp\u003eM, m: Tier 2 classifiers, M: Probable set of candidates, m: selected set of candidates\u003c/p\u003e\n\u003cp\u003eTR\u003csub\u003e1\u003c/sub\u003eC: \u0026nbsp;TR\u003csub\u003e1\u003c/sub\u003eC\u003csub\u003e1\u003c/sub\u003e to TR\u003csub\u003e1\u003c/sub\u003eC\u003csub\u003en\u003c/sub\u003e -Tier 1 Classifiers\u003c/p\u003e\n\u003cp\u003eTR\u003csub\u003e1\u003c/sub\u003eP: TR\u003csub\u003e1\u003c/sub\u003eP\u003csub\u003e1\u0026nbsp;\u003c/sub\u003eto TR\u003csub\u003e1\u003c/sub\u003eP\u003csub\u003en\u003c/sub\u003e -Tier 1 Predictions\u003c/p\u003e\n\u003cp\u003eTR\u003csub\u003e2\u003c/sub\u003eC: \u0026nbsp;TR\u003csub\u003e2\u003c/sub\u003eC\u003csub\u003e2\u003c/sub\u003e to TR\u003csub\u003e2\u003c/sub\u003eC\u003csub\u003em\u003c/sub\u003e - Tier 2 Classifiers\u003c/p\u003e\n\u003cp\u003eTR\u003csub\u003e2\u003c/sub\u003eP: TR\u003csub\u003e2\u003c/sub\u003eP\u003csub\u003e1\u003c/sub\u003e toTR\u003csub\u003e2\u003c/sub\u003eP\u003csub\u003em\u003c/sub\u003e - Tier 2 Predictions\u003c/p\u003e\n\u003cp\u003eWindow: W\u003csub\u003e1\u003c/sub\u003e to Wn (size: 1000 rec.)\u003c/p\u003e\n\u003cp\u003eMetaclassifier: MetaC\u003c/p\u003e\n\u003cp\u003eCC- Component Classifiers - Pool of learned classifiers\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e4.2 Algorithmic pseudocode\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAlgorithm 4.1 is illustrating overall system functioning sequence and accordingly corresponding sub modules are getting called.\u003c/p\u003e\n\u003cp\u003eAlgorithm 4.2 is illustrating building classifier pool initially with historical data and thereafter with recent past sliding window. Part-I is for prequential evaluation and part II is for selecting best performing ones.\u003c/p\u003e\n\u003cp\u003eAlgorithm 4.3 - Part I - Update classifiers online-AMERIT-C -Update/Rebuild\u003c/p\u003e\n\u003cp\u003eAlgorithm 4.3 - Part II - Predict class labels- AMERIT-C_predict\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAlgorithm 4-1: Overall Working- AMERIT-C\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cimg src=\"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAsEAAAFHCAYAAABAsJEsAAAAAXNSR0IArs4c6QAAAARnQU1BAACxjwv8YQUAAAAJcEhZcwAAFiUAABYlAUlSJPAAAJ9NSURBVHhe7P19lFtVeueLf5V7MyvdaWyfkrPodghNdMS948Q91RckQ6DKK2alLLXtcaBhWio6yXgNpI1k4pmAqcIqHO50sPERuLnXIS6VY884BNCpxv3GzxIlO8vMitQO4DKRxnR77qWOQoPj7qyWSrbpdLImN9m/P6y9s8/W0Wup7KrS81nrLFv7fT/7OUePTj3P3i7GGANBEARBEARB9BA/oyYQBEEQBEEQxFKHjGCCIAiCIAii53AtZXeIv/lhCTMf/A2KH1yE9YOLsD74G/zk7/4BP/m7v796/fTqv//4j/+fWpUgCIIgCKIlfvZn/1d86uc/gU998hNX//35T+BTP/9z0G/5ReifXQXPLavgveUX8YufWalWJa4jS8oIfuvd7+P0O9/DX05/D2f/+/+Ln/79P6hFCIIgCIIgrguf/MTP4fZ/87/h13y/irvW/iruvO1X1CLENWRRG8GVSx/jO1Pfxcm/mMZfTn+/xui98Rc0eD579deX55bPwHvLL2LF8k/hUz//Cfz8J39O/GL7V//qZ231CIIgCIIgWuV//s9/FH9h/rufXv2L86XLP7n6V+i/vnj1r9I/uIi//XHFVu+Tn/g5/JrvVzC0zoffDN4NbcUNtnxiflmURnDqz9/Ct9/I4f934rQtffWtn8Vd/l/Fr/l+FXf715AyEQRBEASxYKhc+hjfPfMe/nL6ezh95ns4//4PbPn/dsNd+M3g3dg89Gu2dGJ+WDRGcGn2Mv7k5eP4s9dOYvbSFZH+G+tux71fGMDAHZ/Dp3+hz1aHIAiCIAhiofKjH88i9/Y5fPuNHP78L86K9D5tGX77gSH87m9txsq+5bY6RPdY8Ebw3/ywhD95+Tj+5JXj+Kd/+mcAQP+vevGbwbtx3xcG8Jkb3WoVgiAIgiCIRcUP/7aMb72Rw3emvovC92YAAP/L//Iz+N0vb8bv/tZmCqqbBxasEfzj8iXsH/86jk5OibTNQ7+Gh7+8CXfeTo7kBEEQBEEsTd46+30cfiWF4yf/UqRtDQXxeORL+AX3CltZonMWpBF85NU09v3Rq/j4Jz8FANy/aR1+97c24/NrvGpRgiAIgiCIJUn+vRn8ycvH8Y3UXwAAbvjUJ/Hk7z2Ihx7cqBYlOmBBGcFvnf0+9v3Rq3jr7PcBAJt+407sjIaw+tbPqkUJgiAIgiB6gvPv/wDPH5xE6s/fAgDceduv4MkdD9JfxufIgjGC//CFP8Mf/5dvAQBu+aVP48nfexD3fmFALUYQBEEQBNGTfPuNHPb90av44KMfAQC2/4f7sPv3f1stRrTIdTeC//rDH+KJryaQe/scAOD3Hvoidu34Mn7mZ1xqUYIgCIIgiJ7mn/+Z4dkDr+CPjnwTADBwx+fw3B88gl+++TNqUaIJ19UI/s7Ud/HEfx7HlZ/8FL9882fw3B88goE7PqcWIwiCIAiCICRyb5/DE19N4K8//CGWfeqTeO7pCH4zeLdajGjAdTOCnz3wCv7vP/kGAODeLwwg/gePYNmnPqkWIwiCIAiCIBy48pOfYuSrCXz7jRwA4D/+7v3YtePLajGiDtfFCB79wwn86dczAIA/ePzfI7r1N9UiBEEQBEEQRAscPPodfHX/nwIA/v2XAjB2b1OLEA5ccyN42xP78Z2p78LlcuHw157Apt+4Uy1CEARBEARBtEHqz9/Cw489B8YYfjN4Nyaee1wtQihcMyP4J3/393j4sefw307nsWL5p3DkayO4e+0atRhBEARBEATRAd995z089Fgcly7/BL9+1+dx+GtP4FM//wm1GFHlmhnB4W1fxX87ncctv/RpHP7aE1jzr39ZLUIQBEEQBEHMgff+x1/j4ceewwcf/Qi/ftfnYU78gVqEqPIzasJ8sO2J/cIAfnHPfyQDmCAIgiAIYh5Y869/GS/u+Y+45Zc+jf92Oo9tT+xXixBV5t0IHv3DCXxn6rtYsfxTOPy1J+D7P/53tQhBEARBEATRJXz/x/+Ow197AiuWfwrfmfouRv9wQi1CzLcR/OyBV/CnX8/A5XLhyNdG6A0wQRAEQRDENWDNv/5lHPnaCFwuF/706xk8e+AVtUjPM29G8Hemviv2AT78tScoCI4gCIIgCOIacvfaNTj8tScAAP/3n3wD35n6rlqkp5kXI/ivP/whnvjP40B1H2DaBo0gCIIgCOLas+k37sQfPP7vAQBP/Odx/PWHP1SL9CzzYgQ/8dUErvzkp7j3CwN0EAZBEARBEMR1JLr1N3HvFwZw5Sc/xRNfTajZPUvXjeA/fOHPkHv7HH755s8g/gePqNkEQRAEQRDENSb+B4/gl2/+DHJvn8MfvvBnanZP0lUj+K2z38cf/5dvAQCe+4NHsOxTn1SLEARBEARBENeYZZ/6JJ6rvpz84//yLbx19vtqkZ6jq0bwvj96FQDwew99EQN3fE7NJgiCIAiCIK4TA3d8Dr/30BcByWbrZbpmBB95NY23zn4ft/zSp7Frx5fVbIIgCIIgCOI6s2vHl3HLL30ab539Po68mlaze4quGME/Ll8Svyie/L0H8TM/41KLEARBEARBENeZn/kZF578vQeB6tvgH5cvqUV6hq4YwfvHv46Pf/JTbPqNO3HvFwbUbIIgCIIgCGKBcO8XBrDpN+7Exz/5KfaPf13N7hnmbAT/zQ9LODo5BQDYGQ2p2UQPMzY2Br/frya3RblcxsTEBLxeL3K5nJq9oCgUCggGg3C5XPB6vTBNUy1yTSkWixgbG0NfX5+a1ZRisYi+vj6k09f/T2WLSQcIgiAWC9xmOzo5hb/5YUnN7gnmbAT/ycvHAQD3b1qH1bd+Vs1e0nCDR72i0ahadNGQy+Vq5pPL5RCPx2vS69WJx+O2NufCkSNHsGvXLliWpWYtKMrlMtavXw/DMFAqleD1erF//36Rr8pIvvx+f1dlxonFYti7dy8qlYqatahYLDpAEASxmFh962dx/6Z1gGTL9RxsDvy4fImt6r+f3bjmPvZX595Xs3sGAAwA03VdzbquRCIRNaklSqUS03WdAWD5fF6kJ5NJBoD5fD5beVatA4Alk0k1a86kUikGgGWzWTVrwZBIJFggEFCTawiFQky+7fL5PPP5fAwAi8VitrLdIBAI2Pqrh9OaLiQ60YFUKsUMw1CTu8pClxtBEEQj/urc++zGNfexVf33sx+XL6nZS545vQn+k5eP45/+6Z+xeejX8Pk1XjWbuI7kcjmMj189urpd3G43nnnmGTUZ4XAYgUAAlmWhXC7b8i5cuACfz4dwOGxL7wbLli1TkxYc3/rW1f2xm3HbbbfZPvf392Nq6qo70d69e21514pisYjp6Wk1eUHRiQ4cPz6/bzYWg9wIgiAa8fk1Xmwe+jX80z/9c0++DZ6TEfxnx04CAB7+8iY1i7iO5HI5DA4OqsltMTQ0BAD4+tftDvNbt25FpVLByZNX157z9a9/HQ8//LAtjWgNt9utJl0zyuUyQqGl58tvmmbHPwJbYanKjSCI3oPbcH/2mv17vRfo2Ag+fvIvMVu5gv5f9eLO239Fze55TNO0+X3Kn73ef3lrLvvaer1eRKNRR99atT0A8Hq94jP3QzZN02YAy/0Vi0VRvlgsijJOuN1uhEKhGkPi8uXLAIBvf/vbtvTJyUlhOKNOUJZpmggGg4jH4yLQSZ0nqgZGOBwWY79y5YotH9X2eRmXy4VwOIxCoQAoMuWyckqD5NetjkFGHqvX67WV5b6+mUwGmUympv1W4Guh67qahUKhYJtnNBq1vYWPx+Po6+uDy+VCX18fJiYmbPWbEQwGxdtMl8uFYDAIKGsFZT3L5bKQWzAYRLlcRi6XEzLibcjE43GbDJsF3LWiA/Xmnk6nMTw8DAAYHR2Fq+rX3qhOPdLpNLxeL/r6+lAoFDA2NgbUkZsso0KhAK/XawsMbSQDOahSbg9dkD1BEEQj7rz9V9D/q17MXrqC1J+/pWYvbVT/iFZ5+LHn2I1r7mN//F+/rWb1HPV8gg3DEHmBQIBZliU+y76K3G9T9qlVP7Oqjy9P5/DP3P83m806lmPKeFrxleQ+wKlUSqQFAgHh11oqlRir+rWGQiGp5r/4vvIxZLNZpmkaQ9WnmPt2xmIxBoBZlsVY1bfY5/OxSCTCSqUSsyxLtMXrlEolpmkaSyQSjDHGLMtiPp+PaZom2slmszXy4/JX0xrJwjAM5vP5RLuJRILBwX83EAi05BPM14BV55FMJpmu60zTNJv/NauOLRQK2eakaZpYa74+vB6XEy/PWvQJVsvIa8VlI69nKpVipVKJ5fN5hqpu87XgvruyjGOxmPhcKpWEfsrjlGlFB5rNna+/vLbN6jghr4thGLY1VuWmyiiZTDJN0xhrQQa6rot7iMuVr7PabjuyJwiCaIU//q/fZjeuuY89/NhzataSpvG3Yx1mK1fYjWvuYzeuuY9d/NFVQ6iX4V9QjYxg/mXHPzsZwXJ9/iUpf+k6Gbf8sxwE51SOSUagPJ5G8GA33rZlWSwSiQhjgn/Zyl/wMqqR4GSY8DRu3BiGUSNH3p9sOKtl+NxkOfh8Ppv8+HzktFgsVlcW9YL9VAOGdWAEy5dhGI5j4D8QnC5WNchlOaiyZA5r4IRTGae1cirnNG+5Hpeh01Xvx0crOtBs7k7jb1bHCTVf1q968oD0A5G1KANN02x6psq1Xl+NZE8QBNEqF39UEnbdbOWKmr1k6cgd4jtT3wUA/Ma62/GZG6+fP+Ni5oMPPlCTHMlkMmpSx3g8HlR/+MDj8ajZNXCXCL7f7cmTJ7F582YMDQ1B0zThEqG6QsyFU6dOCfcNzk033WT7fPbs2ZoyHo8HPp/P5ubx+OOPI5PJiLRjx47BMAxbmmVZdWVx/vx5wKH/zZs3AwAuXrxoS28Hvg6BQACvvfaa4xjOnj0LwzBEWfkCgG3btmFmZgaFQgHRaBRbtmxRm7jucBmq42eMYWRkRC0OtKgDncy9kzo+nw+Dg4Pw+/0wTRMHDx5Uizgi+3m3IoPZ2VkMDQ3BNE34/f6u3vcEQRDN+MyNbvzGutsBycbrBToygk/+xVVfODodbulz7733olKpwDRNfOtb38LGjRvhdrsRDocxOTmJXC4Hn893zYO71N0p4BBgxg1zHsRXKBTw0EMPQdM0HDlyBIVCAevXr7fVcUL1R+1kp4J6vPLKK7Asq+7e0qdOnVKTBNw/dHR0FOvWrcPLL7+sFlkwdPuQi07m3kmdqakpGIaBSqWC4eFh+P1+R91rhUYyME0Td9xxB86dO4fDhw8jEAioRQiCIOYVbtNxG68X6MgI/svp7wMABu74nJpFdBn5y/CWW26x5dWjXrl2AuM43JDcv3+/7W0lfxu6detW3HvvvSJ9rqxYsQLvvPNOQ0PD4/FgenpaBMJxZmZmcPvtV3/JQnqTffjwYaTTaTz44IM2A/6tt95q+AZ71apVAICXXnrJln7hwgUAwOrVq23pneB2u/H6669jfHy85oQ5j8eDTCaDaDRqWy++Dd327dtRLpcxNTWFcDjcVeO8W3AZbtmyxRYIlsvl6galtaIDncy9kzq7d+/GyMgIZmZmkEgkMD09Ld7stkozGRQKBQwPD+PAgQPYs2cP+vv7pdoEQRDXBm7TcRuvF2jbCH7r3e/jp3//D1h962fx6V9o/zhWoj6WZYm3RXxXhq1btyqlrlIsFmuMpnrwt4zHjh0TafL/G8ENyenpaWH4AsDGjRuhaRosy6prSHIjhv/L36jyHSbkNP7vo48+ikqlgu3bt6NYLKJcLuPFF18EqrLI5XLYuXMnNE3Dww8/LIzDiYkJzM7O4rHHHhNto/ome3p6Gi+99BIGBq7+yn3wwQdhWRYOHz7s6IbA8Xg8iMVimJycFAZbsVjE/v37YRiGePNcLpdtVyPeffddQHmTPTAwAMMwMDw8jLGxMTEnPs/x8XHoui5+wPAdNy5duoRKpYJyuYxisYg33nhDtMmNLXUNnLjnnnuAahm++4HTWjm1NTMzY/vM/8/rcRlWKhVs2rRJzGFwcLCu3rSiA83mztf68uXLKBQKME2zaR0nxsfHxdrfeeedAIAbbrgBqCM3Jzk3k8HHH38MAPjoo4+A6lhmZmbE/9Gh7AmCINrh07/Qh9W3fhY//ft/wFvv9oghrDoJN2P/+NfZjWvuY7G9f6Jm9Rw8WEW95OAxfum6Lk5h4xcPhOHt6Lpua9MpwEVuQw6Ak9tTy3HaDYzjyFHuMpFIpGZXCI4qGx6EJM9NDRLj802lUmL8gUCApVIp5vP5bPPjO1LwuvIuCjL1gtt0XRdR9c3ggVqorpFcT50XH7NKK+Wc1l4+UY7vDMGDrlKplNjFIRaLsXw+zzRNY4FAgJVKpZo1qAevp+s6y+fzNWM1qrsiqG3Jn+GwxvL8YrGYbXcQeccRJ5rpQLO5MymAke/k0UodFa6nmqbVBK+pcpNlpAb2sSYykJ8B2WxWlE0mk3OWPUEQRKvE9v4Ju3HNfWz/+NfVrCWJi/Eomxb5dw//n8i+/d9x+GtPYPPQr6nZRAcEg0FkMhnoui7eABEEQRAEQVxLjp/8Szz82HMYvOPf4LXD/6eaveRo2x3i7H//fwEAd/vXqFkEQRAEQRDEIoXbdtzWW+q0ZQT/zQ9L+Onf/wNu/AUN2oqrfnEEQRAEQRDE4kdbcQNu/AUNP/37f8Df/LCkZi852jKCZz74GwCA57NXo52JuROPx8WeoJZltbVzA0EQBEEQRDfhNh63+ZYybRnBxQ+uHg7gveUX1SyiQ0ZGRmo20G+0YwFBEARBEMR8wW08bvMtZdoygq0fXBWI/sv0JpggCIIgCGKp4bnlM4Bk8y1l2jOCq6/GdXKHIAiCIAiCWHLwN8Hc5lvKtGUE/+Tv/gEAsGL5p9QsgiCInqJcLmNiYgJer9d2JPLY2Bj8fr+tbCd0q52FSi6XQzQaRTAYVLPqYpomgsEg4vG4muVIuXpU9vWiWCxibGxMHHAzX9TTxfmkWCyir6+v7kEz9ei0XjMmJia63mavwm08bvMtZdo0gv8eAPCpn/+EmkUA8Hq94jQofnm9XrXYkiadTiMajXb00M/lcjXyc7lcarFFDZ+jeuQzJxgMinm3IkO/3y/KB4NBW/16Fzcg6pXt6+tDOByuGaNT+VwuV3fd1KvVL+disSh0yFW9h+LxOMrlsjj9UKXeGOrNpRscOXIEu3btgmVZataiYq461ylbtmwRJ2O2AjeaeSBxMwqFAu644w7s2LGjrn6oV7eJxWLYu3cvKpWKmtVVloouzoUHHngATz/9dN1nBNE63MbjNt+SRj09oxG+DdvYjWvuYz+48CM1q6eRT4eTT5SST2hzOv2tHSKRiJo0L8yln2w2azvdqlP4SXBOJ8AtJFKpVNvrqp5i5kQ+n3fUJxX5lLB8Pi/SeX2nsfET0Dhq2VKpxBKJBEP1hDp1DbhOy3qSTCZtJ9nxcWWzWcaqbYZCoYZz4fB7KRQKiTmVSiVxamEzveK6w+GnxGmaZpNRt0ilUra5dkonutRN5qpzneLz+do63Y7330xWpVJJnOQno+oHx7Isx1P+ugF/Js43reqiz+dTk5YMpVKp5nRRon1+cOFH7MY19zHfhm1q1pKjvTfBP62+Cf4kvQnm5HI5DA8PAwCSySTC4bDI83g8yGazAIDR0VGYpiny2iGXy7X1xqRT5trPwMAApqamEAgE1Ky2uO2224Cq/BYyx48fV5MaUiwWceLECei63lDO/f398Pl80DQNTz31lJotePHFF4Ws+/v7Rbr8f5WRkRHbZ7Ws2+3Gtm3bEIvFUKlUcOzYMVs+X5NbbrlFpH344YfYuXMn3G63VPJfcLvd2LVrFz788EM1ywa/l0KhEEzTFGNzu90Ih8N488031So1cN3hbNy4EQcPHkSlUsGzzz5ry+sGy5YtU5M6ol1d6jZz1blOqaczc+XIkSPYsGFDzRhV/eB4PB585StfUZMXFa3oYrFYxPT0tJq8ZHC73ZicnEQ0GkW5XFaziRbhNh63+ZYy7RnB5A5Rw9atW8X/ZQOYMzAwAF3XAaDhl0s9crkcBgcH1eSuc636WSqYptnQkHXi2LFjeOKJJ/DEE0+gUqk0/FHkdrvx5JNPwrIsx3L8y+yee+5Rs+rCXRdUQ9iJ5cuXq0l1GRkZafqDpb+/v2m/zzzzDABg7969ahZQbSMSiajJTbnpppsAAJcuXVKzFgSd6NJ8MB86dz0ol8vYt28ftm3bpmbVJR6PN9XPxU65XEYoFFKTlxwejwdr167FkSNH1CyiRXrJHaItI/gf//H/AwD8q3/1s2pWT1IsFoUPVqO3n9wv2LIs5HI5mKZZ44cm+xNznybTNG2GqUvyj+RlvV4votGo+CwHjMy1H1TnyMsvxEM8CoUCwuGwbU7yG4B4PC58S/v6+jAxMSHy5CCbsbExuFwupNNpWzoPNpFlm06nxdv/0dFRuFr0dz106BAeeOABPPDAAwCAb3/722oRGw899BA0TcP+/fvVLDz//PPCaGyVF198UU2qywcffAAAuPnmm9WseaFYLCKTycDn8zU0qA8ePKgmNeXKlSsAgBUrVgAN1h0t6FO5XBb5Xq9XtM2pFwglt9vX14exsTGggS7Va6dYLNrGp/o7N9PdZsxF5+T++HNKxTRNUUa+F2Xi8bitnXaDnU6ePIlKpVLzFrgehUIBp06dEp/l56ssO9mvmN/vzdZDpZ2256qLKsFgULwFdlX9uWU9KxQK8Hq9IhizUCjY/MV5eY6sa1B0X5aL1+vtSj0oz/NwOFz3be8999yDQ4cOqclEi3Abj9t8SxrVP6IRN665j9245j41uWeR/eMa+bVxP1DZ305O4/DP3N8ym806lmOSn5ncpvqZdaEfwzBEejM/PM5cfeB4n82wLIuFQiHht5rNZpmmaWJe3L+U+wXKvsa8LKprZ1kW8/l8bOfOnSLd5/MJ/7pYLCbq8r7akQn3m+U083vm+sT7lf38SqUS0zSNsQay4mumXk7+gvI88vm80IV6/uHN5s1l49RXPXidRvdRK3B5yP7JfD2z2WzddU8kEk31ifsbcv9nXl6eK/8MaU0sy2KaprFEIsEYY8Lnmn920iWndvi683p87Nx3W55bI92tx1x0zjAM5vP5RB98jrLvO/fr5u0mEgmm67ptzWOxmHh+lUoloYvt3HeRSKSuHsnPM/lSy/M1U333E4mEzVe90Xpw1OdhK213QxedUMci61kqlRJrxBhjuq6zUCjEmOQzLn9ncF3jayG3ZRiGGJc87k7rseraaVXfflk35LY43D+6mc4T9ekVe6/227MBvSKUVuEPZDg8RGXkm7UdI7heOSY9zORgDl5WHotT/Xb6saTgvlYfKOqDtl2cvmSd4F/WTheTvmQ5fL34lwT/rH4ZOX3R1qurPnzroeu6TX78IV2vPl9DLn95TQ0puK2erJzaDoVCjl+Qqux8Ph9LpVJqMYFT2zKqrFqB12l0H7UClwfXQaf51Fv3ZvpkGEZN8BT/oSXPVdX/SCRiu9e4AdPICGYO7cRisZr+uX7IRobaVqvr0anOlUolBoeAOv5MsSxLGIxqGdkI5u04Xbxvp/mpBAKBunqkjp1VDTyn8rFYrCadrxnPb7YezGEdWYttqzLgF2tDF1WcxsLT+A9Hjrpeqlyd1qJe+92op37m+uI031Z1nqhPr9h7bblDEHZWrWrt0BD5Tzpr16615c0HrW4h1AoejwcL9Tjns2fPwjAMMT75AoBt27ZhZmYGhUIB0WgUW7ZsUZsA2vR/7YRcLgfLsqDruvjT4qZNm4Cqi0QjPB4PIpEIMpmM+DPpoUOH8NBDD6lFm/Loo4+qSQIux2w2O6+BM/KffeU//95www1q0TkxNTUldOHMmTPYuHGjWqRm3Zvp06lTp4SbEIf7GzeiWCxC0zTx2e1248yZM235rKI6PrV/j8cDn89X82fjudCuzp0/fx5wkMXmzZsBABcvXsT58+dRqVRqysjz4e2osmeMzau/bn9/v6Of80MPPYRMJiPcG4rFIn7pl35J5M9lPVppez50sRFqkOLs7CyGhoZgmib8fn9Xv1e6AR/v6dOn1SyCaBkygueAx+MRvsCNHhAzMzMAAF3XF5whudgol8s2X0PZl0+Fb5Q/OjqKdevW4eWXX1aLXBNefPFFZLPZmi8zwzBgWVZTn8edO3cC1XZM08SGDRtqvrBaYWBgAAMDA2qyjYGBARiGgd/6rd9q6NvYbfr7+6HrOjKZTF0/v2tBI32aC5OTk2pSRzjJphNdaEYnOqf6pLayW4ETrfjXdxsnI5sbtNx3+ciRIzU/pjpdj1bani9dbBXTNHHHHXfg3LlzOHz4cMO4l2vBjh07bD/M+PfAXXfdpZQkiNYhI3iOyDs+OEVUm6YpgueOHj0q0uUtphrRajkZ+WHVav165YoLLDDu2LFjIljL4/Egk8kgGo3axsZ36di+fTvK5TKmpqYQDoc7/lKeC8VqRL2T8ckD5Jptj8XfzE1OTuKpp54SBkqn8ODMeoyMjGDt2rV4+OGHHb/k58LAwEDNjwEumwMHDgAAdu/erdT6F8bGxuZND5vp04oVK/DOO++0LZMVK1bAsqyawCbebqt4PB5MT0/X/DiZmZnB7bffbkubK+3oHP+L2EsvvWRLv3DhAgBg9erV4k1/o7d2vJ0tW7bYfhjmcrm6QXROeDyetteIox608Pjjj2N8fLxG5ujCejRrez50sVUKhQKGh4dx4MAB7Nmzp+Ugw/lk48aNiEQieOaZZ+ByuXDo0CGkUinHZyun1b/WEr0LGcFzZGBgAMlkEgAwPDxse4uRU/YQrnezFovFhkaJjPqQ5jtOABDbLMnbtsl00o+8T6y6Z2w9+INZfUCXy2VbZHw93n33XcChvmma2LVrl3Ap2blzJzRNw/j4uM3VgEfUX7p0CZVKBeVyGcViEW+88YZoK51O17y54vD0y5cv16Txf/laXr58GYVCwVGuxWIRoVCo7rZE/G3Q+Pi47Uu+UCigXC7b5v/ggw8CADZs2GD7awIfo/xFyv8vj59jmia2bNmCoaEhoEHZV155BZZl4Y477rAZJPwLme8e4cT3vvc9wOHNYDP4l9z4+HhNlD13abn77rsb/jWlnu7I1BtXM3169NFHUalUsH37dhSLRZTLZbHjxtatW8V9qOr/rl27gOr9uXLlShH5zvesradLajt8fA8//LBYh4mJCczOzuKxxx4DWtRdJ+aicx6PB7FYDJOTk0KPi8Ui9u/fD8Mw4Ha70d/fj0AggH379on5maaJmZkZYezxdiqVCjZt2iTkPzg4KPTVaX4q69atq+vSU08/itVdCWS3FQAYGhqCpmm4//778aUvfcmW18p6wGEdOa20PVddVOFuH+VyWTyH1XEBwMcffwwA+Oijj4Dq85L/RZM/D5zWwmmuql51Wm9iYgKapglXp5mZmZq355zTp0/TX16J1lCdhBvRK47SncKd++VLDX6Q0XVdlJMD06AEmcjlOLwvHljC850CRubSTzuBcTwYQb7k8fDgGDUgieNUX73UYJB8Ps98Ph9D9YSziHRyGT8tDNUgqHw+zzRNY4FAgB0/ftzWphp4I4+fB9PIaayF09/kOk56IOfzMqoOqcEjPNBDHWe9+k4XD9pxKisHkqh9NCvPHOaE9h4xjFX7lSPG+bqqp3/JqGPl41WRy8nrzmmkT6yqU/w+CQQCLJVK2U6oUmXEyWaztnbVflVdqtdOPp+3yUbdQUCu00h3ZdS+2tU5Dg/WQlW2cqAXU3Z74EFXoVCIRSIR27MlFouJ+1YOalT7d5oLk54zjXS53uX0jItEIsxX55S1RuvBHGSr0qztueiiE/wZqFdP05PHpz5beZ6u6yybzYp1SSaTNfI0DMNxrvJnVJ8XndSTx6NegUCgJqgvFArVfS4TrdEr9p6LcU/7Fvj0574IAPjRuW+qWcQ1JhgMIpPJQNd18QudIAiCuOov+sEHH3S0rzSxMDFNU/xlVSWRSIhA02KxCJ/Ph+npaXoTPAd6xd4jdwiCIAhiSfHQQw/hzJkzjm5KxOLkww8/RKlUqokpyGaztp1eYrEYnn32WTKAiZYgI5ggCIJYUrjdbhw+fBjRaJQM4SXAxMQEDh06hLffftvmJ1woFHD69Glxelw4HEZfX1/b2w8SvQsZwYuQeDwutmSzLGvB7NxAEASxUOjv78f777/f9HhyYuHzwAMPIBQKYceOHSK4lO9dzLe3O3bsGNavX08uMERbkE8wQRAEQRAEIegVe4/eBBMEQRAEQRA9BxnBBEEQBEEQRM9BRjBBEARBEATRc5ARTCxqxsbG4Pf71eQ5USwWMTY2Jk5nWgjkcjlEo1EEg0GRViwW0dfXZzvRrRO61c5caPcI4WZ0aw1N00QwGEQ8HlezmtLpnObSZ7s0WvtCoYBgMChOuIvH43XLdpNGY1rqdEtvCYJoDTKCia5imia8Xq/ti7MdcrmcOCK03tVum+0Si8Wwd+9eVCoVNeu6sWXLFnEs9lKjWCxC13VhcHVjrbuxhvyHB9+JpR34nNS0aDSKvr4+2/1RLpfFMeVz6bOblMtlrF+/HoZhoFQqwev14rXXXlOLLWoaPWv8fn/HujcXuqG3BEG0gXqEXCN65Rg9ojNSqRQzDIOVSiVWKpXEka31jjdtBD+KVIUfBTvf8CM6FxI+n8/xKOB2qXdM6/UikUiII5Hz+TyDdLTzXOjGGvJjXtvVOXlOjDGWTCYZqsfq8vRSqcSSyaQ4IpjTaZ/dJJFIdEXXWuF666P6rJGPK74eR+866S1/thLEtaJX7D16E0x0jY8++ggjIyNwu91wu90YGRlBIBDAvn371KJNue2229QkABB7QvYibrdbTWqbYrGI6elpNfm68uabb6K/vx+o7u0KALfccotSanEhzymXy2F4eBihUAimaYp0t9uNcDiMN998U6l9/fnWt76lJs0LC0Ef1WdNf38/pqamAAB79+615V0vjh8/riYRBNEFyAgmuobTKT333HNP1/60x/9k3MuG8Fwol8sIhUJq8nXFyW1gsaPO6ZlnngEaGFT9/f2IRCJq8pJnIeojpxs/OLuFaZpL1hWKIK43ZAQT88rly5e78kVXLpdx4sQJW5pTEIkcVDQxMSH8k1X/Ph7k43K50NfXh4mJCVt+M+S+C4UC/H4/XC4XgsEgCoWCYxmv12sL4ovH4zb/aTUQSPavdhpfvQAquV2/3y/GEwwGxVs3PtZG7cjyU/275bkVi0WEw2FRriidXphOp+H1eoUMxsbGRB4AnDx5El/60pdsaa0iB27x+ch9c3K5nJiHvD6cZuug0s6cisUiMpkMfD4fPB6PrZxMs1Oums210Zga5alrz/1kM5kMMpmM6M+pLKeevqHJuK+HPrYKr+P0A61QKIj2XS4XotGoOMqX+3fz9HQ6jVwuZ/M/5uOOx+MiLZfLKb1cJZ1OY3h4GAAwOjoqytbrhyCINlH9IxrRKz4iRPfQdZ1ls1k1uSncn1i9ZLgvH0/PZrPCv9Ln84l+Y7EYA8Asy2JM8s/kvpm8HZ7P6vjlyfB8ACyRSDBW9SXUdZ3pus6YMr5UKiX8P1l1TMlkkrGqb2gkEqkZo6ZpYg6JRILpui78NOW5yr6CsViM+Xw+ZlkWK5VKTNd10SdzmFe9dgzDEO2wav+QfCTluXE/cMuymKZpNn9eTdOEnA3DqPEzDYVCts/s6gmWLfk/6rou6jv5EvO58vXJZrO29WEtrIOTf247c+L11TKNcOqz2VwbjaleXr21Z1XZyW3UK9tM35qN+1rroxNG9VnDJD9tPg/Zr5sxxizLYqFQyKYfch+JRELM17Is2/MvlUrVzIs/i3gZVkcmar1G/RBEN+gVe6/+t7wDvSIUojskEomOA0vkLyZW/XKSjRdOK18YPI1/SXCDsl4+c2jXCacy6pcaL1MqlUSZUqkkvrDVi3+Ba5omjDOObAQzh7lalsWgGPOy4c3qjFlth49P7V81EJ3aUo0nWRas2gbHsixH/ZDH0ghVRmrfTuPjhkgqlWq6DsxBNqzNOfH68ria4dRns7k2GlOjPKe+mEP7zKFsK/rWbNxOa6T20019dII/a+TLMAzbvDj8B7XTxdsKhULifk8mk0L26rzkNHl91Hk41WvUD0F0g16x98gdgpgXCoUCCoUC9uzZo2Z1hNvt7opbBaq+yzMzMygUCohGo9iyZYtapGOGhoYAAKdPn7alyz6G58+fB65+y9VcIyMjOH/+PCqVCm666SapBcDr9do+q1y8eBEAsHz5cpEWDocxOzsrlWoOH5/a/+bNmwGpn1bw+XwYHByE3++HaZq2P/vPxRUCAGZnZzE0NATTNOH3+1vaVmzjxo0AgPfee6/pOtSjnTndcMMN4v9zodlcG42pUd5caEXfmo27Fbqpj43gax8IBPDaa685uq+cPXsWhmHU6AtjDACwZs0aTE5O4tZbb0U0GsXatWsxMDCgNjNnrlU/BLHUISOY6Drlchmjo6P4wz/8QzVrTnTLoC6XywgGgxgdHcW6devw8ssvq0U6hhu7N998s5pVw3z58J08eVJN6ogrV67YPi9btsz2uRWmpqZgGAYqlQqGh4fh9/uF/6S8g0KrxKv76qLqO3rHHXfg3LlzOHz4MAKBgFq8Bl5XXp9216GdOfX390PXdWQyGVGmE5rNtdGYGuV1g0b61mzc7dANfWyFV155BZZliUBclVOnTqlJgo0bNyKbzWLDhg0YHx+HruswTVMtNmeuVT8EsdQhI5joKuVyGbt378Yrr7wiDEI1GOd6s337dpTLZUxNTSEcDnf1y5QHVa1du1bNEqxatQqoHoAhB2HlcjlMTEyIt4fq2+Rm8HrRaNRm2KXT6ba+IPn4XnrpJVv6hQsXAACrV6+2pTdi9+7dGBkZwczMDBKJBKanp3H+/PmaHRRaoVwu49SpU3C73SgUChgeHsaBAwewZ8+elo1pbrCtXbu26TrUo905HThwQNSrx9jYWN0ArlbmWm9MzfLmQjN9a2XcrdBNfWwFt9uN119/HePj4zX3jcfjQSaTQTQata0XPx0wHo9j1apVME0TlmXB5/Ph6NGjQJfH2agfgiBah4xgomvwSPDx8XGsXLlSRD5//vOfx+c+9zmgasj09fU1NYrfffddQIrSrocclQ3pbdHly5dFGZ7G/7106RIqlQrK5TKKxSLeeOMNUZYbQ2q7jeAGU7FYxNNPP41YLCb+lOpU3+PxIBaLoVKpYNOmTUJOg4ODGBoaQn9/v9hfmX8Jm6aJmZkZ8QUMh7n29/cjFAqhUqlgcHAQrurOFzt27BBuGvfccw9QHRdfA7UdPr7JyUnb3Pbv3w/DMMSPGycZlctl2+fx8XHRxp133glUjSfVbYDDdxaQ1w/V/rdv3y7k+vHHHwPVvalRXbeZmRnxf1TngarsUG07Go0imUzC4/E0XQc4yAYdzGnjxo2IRCIYHx9HOByu2T0hGo3i7rvvFuNV+2xlrvXG1CxP7QvSGqprqZZtpm+tjPta66MT/FkjlxsYGIBhGBgeHrb9QNm5cyc0TRNvX7nOyDvUhEIhFItFeDwe6LqOFStWANJfiXh/uVwOr776KgBg69atQk/VeXA3h8uXL6NQKIhy9fohCKINVCfhRvSKozTRPjwSWw0W4RcP4OBBX04BUUwKApGveoEtPICEX2pdo3q6nJqWSqXEWGOxGMvn80zTNBYIBFipVKpptx68HA+WUeclt+MU1BeLxcQ4fD4fS6VSIk/eqYAHF4VCIRaJRJhlWY5z5fXkdgOBgC3Cnc9V13WWz+frtsOqwTe6rovx810WmIPs2VWnyJo0vgaaptmCpJx2hVDbdLrUICs+tmw2K+Ytl5HnoMqYU28d6smmnTnJZLNZ2y4GWnVXAXl96vXZbK71xtQoz6kvNQ1VHVLTW9W3ZuO+1vooo/bFxy8jt8vHkpdOlONryJ9vhmGwfD4v6vFnCocHDmqaJuSt6zpLJpMNnz38WcCfL836IYi50iv2notxj/4W+PTnvggA+NG5b6pZBNFzBINBZDIZERRDEARBEEuBXrH3yB2CIAiCIAiC6DnICCaIDlF99wiCIAiCWDyQEUwQHSAf+bpy5Uo1myAIgiCIBQ4ZwQTRAVNTUzUb5RMEQRAEsXggI5ggCIIgCILoOcgIJgiCIAiCIHoOMoIJgiAIgiCInoOMYOKaUy6XMTExAa/XaztulVhcFItF9PX12Y4cJgiCIIjFAhnBXaBYLGJiYgJ+vx/xeFzNbhluGLpcLvj9fuRyOZimueQMxSNHjmDXrl2wLEvNIgiCIAiCuCaQEdwFnn/+eTz33HNiy6xOSKfT2LVrF44ePQrGGCYnJ/Hiiy9ieHhYLQq/368mLSpGRkbw8ssvq8lNSafTc/qRQXQXj8eD2dlZbNy4Uc0iCIIgiAUPGcFd4ODBgzh69Kia3BYHDhzA2rVrMTAwAFQNDNM0EQgEbOWKxeKcjO2FwrJly9Skphw/flxNIgiCIAiC6AgyghcQMzMzNaePbd26Vfy/XC4jFArZ8nsF0zQxPj6uJhMEQRAEQXQEGcELhHvuuQeWZeHWW2+FaZoiPRwOi7fD8illLpcLwWAQxWIRY2Nj6OvrQ6FQgNfrtblLxONx4Wfs9XptQUyFQgHBYBAul8vWHqpvnHm75XJZlAsGgyiXy8jlcqLdYDAo2qxHuVxGOBwW47hy5YpaBPF4HH19fXC5XOjr68PExARQdYPgbiGjo6NwuVzCT7peHSdUWfn9fjH+QqHgWKYdeTqhttdJn7KvuNfrdXQJkcfl9/tF20756rjT6TS8Xq/of2xsrKU80zQRDAbFeOR5FItF23pzveLI6xYOh2t+/BEEQRDEvMPa4MY197Eb19ynJhOMsWw2ywAwwzDUrJaJxWIMAAPAdF1nyWRSLcICgQCTly0UCok6qVSKJZNJpmkaY9X2eBulUolFIhEGgFmWxRhjTNd1FgqFGGOM5fN5BoBFIhHHdkulkigTCARYIpFgjDGWSqUYAMexckqlEvP5fCwSibBSqcQsyxLtZ7NZxhhjyWSSAWD5fJ4xqX8+Vif5NqujwmUHQIw/n88zXdeZruuMOcy7HXk6Mdc+DcNgPp9P9JFIJBgAFovFRB+xWEyUKZVKTNd1UZ/nNxq3pmlChoZhsEAgIOrWy8tms0zTNNuayPMwDEOstaZpQq94O7xdeTxyWwRBEMT1o1fsPTKCu4STkdYJ2WzWZjgFAgFWKpVEvmoEy2lyuVKpJNpQLz5GTdNsxmsgELAZQPX6ksuwq2cGN5y3YRjC4ONwA5YbwYlEwlaGy5PnO8m3WR0nnOakjqVTedZjrn2qPzBkI9ayLPF/jmxEtzJueRy8fU6jPKc1cZqrqjPqZz7GRutGEARBXDt6xd4jd4gFxsDAAKamppDNZuHz+ZDJZLB79261mCNut1v8//z588BVa6TmGhkZAQDMzs5iaGgIpmnC7/cjk8mI+t3k1KlT8Hq9trSbbrrJ9nnbtm2YmZlBoVBANBrFli1bbPlOdFLHiaGhIQDA6dOnbentyDOXywm3En412tqunT5VWW3evBkAcPHiRVy8eBEAsHz5cpEfDocxOzsLtDBuAPD5fBgcHITf74dpmjh48KBoq1Fet+BzVmVBEARBEPMJGcELhGg0avvMjWFd120+wu3SyBAzTRN33HEHzp07h8OHD9fsRHEt4X7Ho6OjWLduXUtbqHVSxwluhN18881qVg2N5NkO7fSp+k877axx8uRJNclGo3FPTU3BMAxUKhUMDw/D7/cLH91GeZ2yY8cOZDIZm183ANx1111KSYIgCIKYP8gIXiAUi8UaQ8XtdiMUCqGvr8+W3gqrVq0CAGzZssUWBJXL5TAxMYFCoYDh4WEcOHAAe/bsQX9/v1S7u6xYsQLvvPNOQ+Np+/btKJfLmJqaQjgcdjT0VDqp4wSXz9q1a9UsQTN5DgwM1Lxp5QGNTrTT50svvWRLv3DhAgBg9erVuOGGG4DqjyhZf9LpNEzTbDpuANi9ezdGRkYwMzODRCKB6elp8Qa5UV6nbNy4EZFIBM888wxcLhcOHTqEVCrVUF4EQRAE0W3ICO4S/G3d5cuX1SwRCa9GyKts2bIFpmkKYzGdTmN8fBzPPPOMKHPPPfcA1begPFLfybj0eDyIxWKoVCrYtGmT+BP94OAghoaG8PHHHwMAPvroI6Da18zMjPg/pHbl9tVt3Pj/nebNefTRR1GpVLB9+3YUi0WUy2W8+OKLQHULuFwuh0uXLqFSqaBcLqNYLOKNN94Q9dPptDCQLl++jEKhANM0m9ZpBDcAi8Uinn76acRiMXg8HqBDebZCp31OTk7a6u7fvx+GYcDtdqO/vx+hUAiVSgWDg4NwVXfJ2LFjB4aGhloa9/j4uGj/zjvvBABhXDfKc9J5J50pl8u2zxMTE9A0DVNTU2CMYWZmhg7cIAiCIK49qpNwI3rFUbpdeDCQfMnwaPhGuwjEYjFWKpVEWVR3iEilUrZy+XyeaZrGdF1n+Xze1rcafMaq7fL2fD6frT1eV9d1ls1mRdlkMlkzJ8auBknJFw+M4pcaMCeTSqWYruuiXCqVYj6fTwR9pVIpMc5YLCbmGZACA3lAGN8ZoZU6KnxefCcOTdNsOy3MRZ71mGufPLCQ5/NdJjilUsk2rkAgIHZ04DQat2EYQu/4+jfLU9feqO4cIacxB51hde4XPu5660YQBEFcO3rF3nOxq19ULfHpz30RAPCjc99UswhiURAMBpHJZNCG2s+Z69HnQsY0TcfjwAEgkUhg27ZtajJBEARxDekVe4/cIQiCuKZ8+OGHKJVKNT7U2WzWtssFQRAEQcwnZAQTPYWTz+p8cz36XKhMTEzg0KFDePvtt23yKBQKOH36NMLhsK08QRAEQcwXZAQTPYN87PTKlSvV7HnhevS5kHnggQcQCoWwY8cOrFy5UhzznMlkxL7FBEEQBHEtIJ9ggiAIgiAIQtAr9h69CSYIgiAIgiB6DjKCCYIgCIIgiJ6DjOAeo1AoIBgMwuVywev1zulI5vmAHyzicrkQDocpmKxFisUi+vr6mh4Scq1YaOMhCIIgCBUygudIuVxGOBwWJ3GFw2EUCgW1WEO4Uapefr8f8XhcLd4x5XIZ69evh2EYKJVK8Hq92L9/v1rsuhGPx/Huu+9idnYWyWQSJ06cwNtvv60WIwiCIAiCmDNkBM+R3bt349FHHxX7nFqWhfXr17f1BnNqagr5fB4AYBgGGGOwLAt+vx+jo6OIRqNqlY44duwY1q5di/7+frjdbkxNTeHMmTNqsevGvn378OijjwIAwuEwZmdn53ycbjqd7uoPiYWKx+Ppiry6xUIbD0EQBEGokBE8B3K5HLZt24aBgQEAwMDAAF544QVUKhUcO3ZMLd6Q/v5+22ePx4ODBw9C13WMj4+3/XbZiW9961tq0oIhl8uhUqmoyXPm+PHjahJBEARBEAQZwXNhYGCgxnjlBvHly5dt6Z3i9XoBAB9//LGaRTTBNE2Mj4+ryQRBEARBEGQEdxvuBhEIBNSsjuDt3XDDDSItHo/D6/WK4DYefFQsFjE2Noa+vj4UCgV4vV74/X7kcjm4XC5kMhlkMhnhc8yZmJiwtSe7D9RrU04vl8vCrzkYDKJcLiOXy4k2g8GgaM+JYDCIwcFBAMDg4CBcLpcYQ6FQsPlcR6NRm6uJHEjX19eHiYkJoOoGMTw8DAAYHR2Fy+XCwYMHRTu8/Xg8LtJyuRxQNZ6DwSDi8TjGxsbgcrmEjJuNx4lW1qtYLIp2vV4visViTRuNAgblMaODtp3G1yy/nm6gC+NpNFeCIAiC6AqsDW5ccx+7cc19ajIhkUwmWSAQUJNbAgCLxWKMMcZKpRKLRCIMgK29WCzGkslkTRnLslgoFGIAGACWSqVYMplkmqaJuoFAoGZshmEwn8/HLMtijDGWSCRs46jXpppeKpVYPp8X400kEowxxlKpFAMgxlyPbDbLALBsNivS+Jz42LLZLNM0jUUiEcaqsgbA8vk8Y9JY5fIAmGEYok0+HjmNt5PNZkUffB6WZTGfz8cSiUTT8TjR6noZhsFKpRKzLKumTcMwmKZpLJ/P29rg9eQx83m12naj8TXLV3WA68ZcxtNsrgRBEMT80yv2HhnBXaRUKtkMynYBwHRdZ7quMwDCQCiVSoxV2+cGgXpxAyEQCDAAoo6MagTz9lQDVTWE6rXJ09U01dBuxYBxMoJjsVjNPPnFqga7rut123AyghulqfX4DwFOs/GotLNeMqoM1c+8XVlWTvNq1naz8TXLl/tQdaOT8Th9dporQRAEMb/0ir1H7hBdZPfu3Th8+DA8Ho+a1TJf+cpXMDMzA8YYZmdncfDgQbjdbgDA+fPngatWRM01MjJia4fXaQRv76abbrKlb968GQBw8eJFW3orbXaTs2fPit0y1AsAtm3bhpmZGRQKBUSjUWzZskVtYk4sX77c9rnZeFTaWa924Otw+vRpNastmo2vWb7MfOlGt+ZKEARBECpkBHeJeDyOdevW1QTKzQfcd7VbXLlyxfZ52bJlts/Xk1OnTqlJAu6LPDo6inXr1uHll19Wi3SdRuOpx1zXa8eOHchkMqId7md71113KSU7o9n4muV3k/meK0EQBEFwyAjuAqZp4uabb0Y4HBZp0Wi0JuBnrqxatQoAsGXLFlvwUi6XEwFh7cDbe+mll2zpFy5cAACsXr3aln6t8Xg8yGQyNbLkct6+fTvK5TKmpqYQDodbMt7nMqdm41Hp1npt3LgRkUgEzzzzDFwuFw4dOoRUKiV2IumUZuNrlj8fzNdcCYIgCEKFjOA5Eo/HMTw8jOHhYbFjgMvlwokTJ4RbBI92b2QU832AG22t5vF4EIvFUKlUsGnTJtHX4OAghoaGAGk3CZVyuWy7ILU3OTkpjJpisYj9+/fDMAzxp+hGbcr/AsDMzIztM/9/o3lBMrz5vwCwc+dOaJqG8fFx6Lou5tvX1wcAuHTpEiqVCsrlMorFIt544w1RN51O27arKxQKME1TzOndd98Fqgbdq6++CgDYunUrTNOseTPOaTYelXbWS5WZ/HliYgKapmFqagqMMczMzNQcQsHHLMu5WdvNxtcsX21bppPxoMW5EgRBEERXUJ2EG9ErjtKt0ihQSg6q4hHv9QLmeMCQfDUKBIrFYiL63ufzsVQqxZjSjlPAmHzJwUeGYYhgPF3Xxc4OrEGb6pgZuxoAJ19qv3KfMmpbvC5jjOXzeebz+RgcAgVTqZSQQywWY/l8nmmaxgKBgCjDg/zk9eC7GGiaxozq7gq6rrNkMsmOHz8uxqDrek1AX6Px1KOV9UIdGTqV4xefpypnwzBq6tRru9H4muXX0425jEctxy95TQmCIIj5pVfsPRe7+mXUEp/+3BcBAD869001iyCIecI0TbHnsUoikcC2bdvU5EVLL82VIAhiodIr9h65QxDEAufDDz9EqVSq2aEhm83W7GCx2OmluRIEQRDXFzKCCWIBMzExgUOHDuHtt9+2+c4WCgWcPn26blDeYqSX5koQBEFcf8gIJogFzAMPPIBQKIQdO3Zg5cqVcLlc8Pv9yGQyNXv1LnZ6aa4EQRDE9Yd8ggmCIAiCIAhBr9h79CaYIAiCIAiC6DnICCYIgiAIgiB6DjKCCYIgCIIgiJ6DjGCCaJNyuYyJiQl4vV7kcjk1uyWKxSL6+vpsxxETi5v5XlPTNBEMBhGPx9Ws60I37oPFgN/vx9jYmPicy+UQjUYRDAZt5erRjpxa7Ustd71Ip9Pwer0iiJWffHo9KBaLGBsbqznBs1uy6lY7xMKCjOAuYJom+vr64HK5EAwG6x4l2wz+oOQPlFwuB9M0mz44iWvLkSNHsGvXLliWpWYRxLzAjaFMJqNmXTd69T7YsmULxsfH1eS6zEVO7fZ1LcnlctixYwfefvtt5PN5VCoVfP3rX1eLXTNisRj27t2LSqWiZhFEfdQj5BrRK8fotUMqlRLHDPPjd+sdD9wIfgQwPy7YsiwWCoUYmhyhvNDw+Xxq0pIklUpdt7XploxTqVTNsdCLiW7JYbHAj6NeSGt2Pe+D6wk/yrtVDMPoWE6t9hWLxTpqv1NCodCC0kVWlUGbZo0j11qWC5FesffoTfAcuXLlijjKdWBgAF/5ylfwzjvvqMWacuDAAaxduxYDAwMAAI/HA9M0EQgE1KILlmKxiOnpaTV5SbJs2TI16ZrQTRkfP35cTVo0dFMOROdcr/uAqGVyclJNmleudX+t0K1TJRfi3Ij5gYzgOaKeYvXBBx8gEonY0lplZmamxpVi69atts8LlXK5jFAopCYTXaSbMjZNc8H+mbUZ3ZQDQSwFotFoR+4WRC0ky96CjOAuUSwWEY1GAQB79uxRs5tyzz33wLIs3HrrrTBNU6SHw2HxdpgTj8eF77DX6xWBOHJgQKFQgN/vF37KcsBCoVBAMBiEy+US+cVi0bENr9cLv9/ftF4wGBRv5ngep954nZCDf8bGxuByuUT5QqGAcDgs+o9GozXH6/L8vr6+miCGYrFoqx8Oh21yadR3uVwWdb1eL65cuSK1fDU/Go2KcaXT6Ya+3GqQkyx3eZxer7clGcv+5F6vt2HwVDqdxvDwMABgdHRUtMXl4nK5gOq6qWmojsPlctnG3kiu9ZD1Qg2qmS9di8fjwn8/HA7X/OhEE1k2uj/UNeU0GhMPLOLtqTrrxOXLl4V+9vX1OfbH59jX14eJiQmR10xPm91jze6DerQq03q6X496c1X1mc9R1mko8uB9yvKoF5im0qwdzoULF8TYVF1otS+13NjYmPhBOzg4CFed+xnK/J3Gh+pc+Hq4qvemPE7eBpTnh4ocEJhOp23rKj9X5TLBYFDoAar3B/8ec9J1+XtOfSEFB1lx5Ha9Xq/4znWSZaN2crmcbXzy/aLOjcu0bx6DZ4kOUP0jGtErPiLtwn31ADBd1zv2JeL+TLydZDKpFmGxWEykl0olFolEGABmWRYLBAKiPvdTzufzTNd1puu6aEPXdRYKhUQ+ABaJRBir+nnxNlKpFEsmk0zTtKb1GGOif5lG41XJZrNM0zThA2dZFvP5fCyRSAgfaV6Pl+X9W5bFNE0T804kEjY5lEolWz5vW9M0ZllWw75LpRLz+XwsEomwUqnk6K+dSCSEbCzLaqgHcl/cp06Wu2EYoh95jqyOjA3DYD6fT8iGzz0Wi9nKyTj5l/I0We8sy3JM4/WaybUesVhMjLlUKjFd14WesXnSNcMwmKZpLJ/P2/JlOTSTZb37w2lNWxkTHw/vu5HvJ18fWW78mcH7SCaTDIBok49Xnk89PeV6zcuq91gr94ET7ci0ke6rNJsr91dOpVK2erFYjJVKJcYYY5FIhOm6zkqlktBD+VkpPxNk1LRm7XCfYFl2XIdlXVDbZQ59OZVTfY5lXePy4ei6XpMmEwgExDjldlQ58vWqBx+TPO9SqWSbN78nUdWHfD4v7olUKmVbK643XNd5WTkmx+fzMUjPBSdZqfE3XG/k+1CWZb12+P3By/HPPFZBnhu/r/j8Zd1YqPSKvUdGcBfhX4haEwOgEdlsVjwk+E3HHwKlUkmkqxd/GDkZB/zLgt+smqbZjJpAIGC7uXkbvF9Oq/U4rYxXhX/Rqwac/ANBvVj1S0j+wixVv7D5AzIWi9U8eKyqgcfr1evbMIyauqpMDcNgoVBIyCyZTDY0DHhfshxU+fG0VmQsrwurygN1fmywOv2zarCZ3B9vX06LxWKi3VbkqsLz5bHxe4czH7qmtsHryIZDK7Ksd3+oMm1lTHL/rNpXPdT2OT6fT3zxJhIJ23rwOq3oabN7rJX7QKVdmcqo66XSbK6sKhtZXqVSyTaWUChkkzk3gGTQgmHarB3+WR4bl438vFHbbTWtXvua8kMimUw21DH+w0F9bqhGPWvBCGZ1xsXvf16Xl1GNbG7Qqhefd6BqrMtwvZJRZaUrL5hkw5vVGTNzaIf/IJTh9wNv36ktVTcWKr1i75E7RBcJh8N4/fXXUalUcPLkSTW7JQYGBjA1NYVsNgufz4dMJoPdu3cDAM6fPw9cvXtqrpGREaWlf2FoaAgAcPr0aQDA7OwshoaGYJom/H5/3W2X3G637XOr9TidjhcOAQ5nz56FYRg17Vx9Nl39k6qmaaK82+3GmTNnRNDi2bNn4fV6RT6qwYc+n6/mT65q36dOnaqpe9NNN9k+r1mzBpOTk7j11lsRjUZtQY7zCZexOp7NmzcDAC5evGhLb8bjjz+OTCYjZHLs2DEYhmFLsywLHo8HaFOuHD4mWc7hcBizs7Pi87XQNa7f/L5oV5bq/aHSyph8Ph8GBwfh9/thmiYOHjyotNKcf/fv/p1wD9m2bRtmZmZQKBQQjUaxZcsWW9lGetrsHmvlPlBpV6bt0GyuqOrzvn37xOeTJ0+K5yGqLiwHDx4Uf64fHR0Vee3QSTtutxuBQABnz55Vs7qC2+3Gk08+ifHxcXEvHj16FA8++KBaVPDee+8B1XtYJhQKdc1Plj8fTp06ZUtXgyynp6eRzWZrdHFqagoAkMlkcMstt9jqqJ+dsCzL1ld/fz9mZ2fR399vK9eM6enpmv64bn344Ye2dGLhQkZwl+nU8IlW/Yk53BjWdd3mI4yqH1I78C/rm2++Gag+sO+44w6cO3cOhw8fbnkHik7rtTveeqgPTZVmEb1O/p/NDJlW2bhxI7LZLDZs2IDx8XHHdZtPVN9M9QulVfhDnP+IKxQKeOihh6BpGo4cOYJCoYD169fb6nQq10Y/FOdD13bs2IFMJmPzDwWAu+66y1auW7LkNBrT1NQUDMNApVLB8PAw/H6/ozwbcfnyZei6DlTXIhgMYnR0FOvWrcPLL79sK9tMT5vdY53SbZmihbmiqs+VSkXM8dy5czbd5H7dx48fx+/8zu/AMAypdut02k65XK4xOLsJv3eff/55FItFzMzMtPQdpeqg+mJgrrjd7pbmzX+gdptu7Yxz+fJl2+dWnnvEwoKM4C7DHx7y24ZWKBaLNV+WbrcboVAIfdUTcFatWgVUN1BXAyrk4BcVXnbt2rUoFAoYHh7GgQMHsGfPnpZ//XZSr9PxOuHxeJDJZBCNRm1vGHkwxIoVK2BZVk0gD8/3eDyYnp6uCdiamZnB7bffbktTWbFiBd55552aLwaZeDyOVatWwTRNWJYFn8+Ho0ePqsW6DpfxSy+9ZEu/cOECAGD16tW29GZwnTt8+DDS6TQefPBBuN1uhMNhTE5O4q233rLpdidyveGGG4DqDz9Z59PpNEzTnDdd27hxIyKRCJ555hm4XC4cOnQIqVRKGAXdlmUrY9q9ezdGRkYwMzODRCKB6elp8fa0VSYnJ7FhwwYAwPbt21EulzE1NYVwOFxjbDbS01busWb3gUq3ZSrTbK6o6nMkEsFTTz2FdDqNu+++25a/fv16hEIhHDx4EBs3brTltUMn7RSr2/ytW7dOzeoa8tvgWCyGJ554Qi1ig78oOXbsmC39gw8+gM/ns6V1SrlcRiaTaTpvXdcxOjqKiYkJW7AZf2GkaVpHP9o0TcP4+LjtO4gHZ7aDruuYnJy03Q/8vlmzZo1UkljQqP4RjegVH5FW4QE7RjWYo1R1enfyKW3mJxwIBIQfJPfX4w78sv9SPb891bdODVTiY+J+czw/lUoxvXrAB/fJUoMLWq3HfZ1K1YAd1sJ4VbhPmuprZlUDZdR2uE8WXwv50nVdtMPrqwE6mqbZ5O3UN5+7HNzAgyn0amCRoQT/hEIhEYDkBO9L1hUud9nXVPb3ZE1krK65Og8V3n8+n7fpGPdrk8fPZSCPhbUoVyfkYCgATNM0plcDi+ZL1xKJRM29qdKKLJ3uD1ZnTZuNSe6L63C9oCXuuxioBm6yqg+kz+cTsg5Ug25K1eAr3n82m2Wp6uEo9fS02T3Wyn3gRDsybaT7Ks3myqmnu6wqfz7/fD4vnp+WZbF8NYBSreuU1qwdfk/FqoFeJSnIsFG7rabxOaaqhzfx9eXltWqsSqN7kuOrBrby9cwqAWCsBb9/Dr9PI0qgnTx2ed1k+P2kXqq/Lf/+zefzQo+4/7KTrBLVADv5kvXXSZZO7fDxqXOT/Yb53GR95GmtrMX1pFfsvdoneQN6RSitUpIiXVF1mpeVndOKEcwfjrwsvzGd2ovFYqKMz+ezleHj4Teapmk1X/y8DL/xeXvJZNI2HzUQolE9Jn1J60r0caPxyvCHD+9DNeLkh5xWDfiQHyRZKTpY0zTH+rLxxb/Med1GfXNDDNI6+3w+2wNZ/vILSAGNKnJfqD7EZbmjamDJn3laPRkb1aAlPn5ucDSCB5Go+sEf+GowU712G8m1HqWq8cr1IhAI2OYzH7qmyphf6lo1kmW9+8NpTTmNxmQYhrjn5fnVQ5a10z2Qqv5whhJtz+fYTE+b3WPN7oN6tCpTNNB9lWZzldF1vUbPmWRIadVdBrhhw8vKYwhUjRuntGbtsOp4ZSNNfcY4tdtqWqlqVNfToVgs1tRg5XBjTtbZrENgl3ypc+HwsnJ7ISkwU26L3+syyWTSpjdqP/zeQVXWRjXwk7fjJCsmBVXyduV70kmW9dqR11Srftc6zQ1VI1/VdXW+C4lesfdc7OoCt8SnP/dFAMCPzn1TzSIWCMFgEJlMRgSzEARxFdM0xf7IKolEQgRREsRSw+/34/Dhwy25FnWTeDyO0dFRZLPZlnyRiYVDr9h75BNMEERP8OGHH6JUKtVEm2ez2a4H/hDEQoEHBV5rA5ggFgNkBC8x5AACgiCuMjExgUOHDuHtt9+23RuFQgGnT592PG2KIBYrhUJBnAw3PDyMF154QS1yTeC7J6i7gxDEQoGM4CWEfJzsypUr1WyC6FkeeOABhEIh7NixAytXrhTHwWYymbr7CBPEYuWGG26ApmnQdR3JZPK6uCLE43Hs3bsXALBp06aa3Y8IYiFAPsEEQRAEQRCEoFfsPXoTTBAEQRAEQfQcZAQTBEEQBEEQPQcZwQRBEARBEETPQUbwIiYej6Ovrw8ulwvhcLjnd4Qol8uYmJiA1+ulIAyCIAiCIBpCRvAiJR6P491338Xs7CySySROnDiBt99+Wy3WUxw5cgS7du2CZVlqFkEQBEEQhA0yghcp+/btw6OPPgoACIfDmJ2dxcaNG9VibTE2Nrao36COjIzgySefVJMJgiAIgiBqICN4EZLL5VCpVNTkOTM5OakmEQRBEARBLEnICCYAANFolNwICIIgCILoGcgIXmQEg0EMDg4CAAYHB+FyuRCPx4HqUZnhcFgclxmNRm3BcnIgXV9fHyYmJoCqG8T4+HhNm7wd7iIRDAZFGpRAtHQ6jWAwiL6+PhSLRQBAOp2G3+8X/fFxquRyOdGuy+VCMBgUeTyN102n0/B6vbY5NkJtU+5LHk+5XEY0GhXyCQaDKBQKUksEQRAEQSwpWBvcuOY+duOa+9Rk4hqTzWYZAJbNZkWaZVksFAoxy7JEGU3TWCQSYYwxlkwmGQCWz+cZY4yFQiEGQJQ3DKOmzVgsVpMWiUQYVxvDMJimaQwAi8ViLJ/PM03TWD6fZ6lUisViMVYqlRhjjCUSCQaAJZNJ0ZZMqVRigUCA+Xw+W3o+n2eJRIKx6hwBiM9ObarzKJVKDAALBAKiDG/HMAyRFgqFRB3LspjP52O6rot8giAIgugVesXeozfBS4QjR45gcnISuq7D5XJhcHAQlUpFvOG9fPkydF1Hf38/AIiguosXL9rakVm+fLmahFtuuUX8Xw5Eu/vuu9Hf34/Z2Vn09/fj6aefxt69e7Fy5Uq4XC488sgjAICjR4+K+jJutxs7duyAZVm2t9dvvfUWtm3bBgD4+OOPAQBDQ0MAINI//PBDUV7F7XarSfB4PLbP6XQak5OT4i24ruuYnp6GZVmLOlCQIAiCIIj6kBG8RDh79iwMwwBjrOZC1WCcmZlBoVBANBrFli1b1CbmxLJly2yfp6enkc1ma8YyNTVlKyezceNG9PX14dixY2oWAKC/v1/MJx6Pw+v1qkU64r333kMgEKgZK2MMAwMDanGCIAiCIJYAZAQvIU6dOqUmCcrlMoLBIEZHR7Fu3Tq8/PLLapGuc/r0aTWpKaFQCM899xwAwDRN3Hnnnbb8sbExbNiwAcuXL8eJEydseXPhnXfe6fnDRgiCIAiilyAjeIng8XiQyWQQjUZFYBqqewgDwPbt21EulzE1NYVwOFzz5taJNWvWqEkto+s6RkdHMTExIYxLHnzWiIceegiWZcE0TfzFX/yFcN8AgImJCezduxdvv/02tm3bVuPWUA+fz6cm2bj55ptRqVRqguHi8TgFxxEEQRDEEoWM4EXIhQsXbP8CwM6dO6FpGsbHx4VfsKu6KwMAXLp0CZVKBeVyGcViEW+88Yaom06ncddddwEArly5gomJCRSLRWEof+973wOqb2b522av14tisYjLly+LdmQOHDgAAHjkkUeEX/DKlSuxbt06tagNj8eDQCCAp556ymYAo+rXjOq8y+UyTNMUeel02lbmypUrIs/tdmNmZkbMne8KsW/fPoyNjSEcDsPn82F6ehqf//znhewOHTpUMwaCIAiCIJYIaqRcI3olWnAhEwgEGADbxXc1yOfzzOfzMQBiZwi+O0MqlXLcySEQCLBSqcRKpRLz+XxM0zTbbgt8hwhd11kqlWKGYbBAICD+z8eg67ptFwlW3ZFC13WRL+/G0Ai+kwUfO8eyLNFeIBCw7eKQz+dt41HlwutxmWiaxgzDELtjlEolsWMGABYKhcROGgRBEATRS/SKvediPNKoBT79uS8CAH507ptqFkEQBEEQBLEE6BV7j9whCIIgCIIgiJ6DjGCCIAiCIAii5yAjmCAIgiAIgug5yAgmCIIgCIIgeg4yggmCIAiCIIieg4xggiAIgiAIoucgI5ggCIIgCILoOcgI7kHGxsbg9/vV5EWHaZoIBoPiBLh6ac1ot065XEYwGFSTiQZMTEyIU/0IgiAIYiFARnCXGRsba9tACgaD4qhe9fL7/YhGo8jlcmq160q9Mfv9/paNybmQy+UQjUaRyWQapjWj3TqFQgF33HEHduzYoWYtCNR1GRsbU4s4Ui6XMTY2hr6+PrhcLkSjUbVIyxSLRUSjUdEWADzwwAN4+umn59QuQRAEQXQTMoK7SC6Xw969e9XkpkxNTSGfzwMAIpEIGGNgjMGyLDz++OMwTRODg4MtGzTN2LNnD86cOaMmt4U8ZsMwxHj9fj9GR0fn3dgZGBjA66+/3jStGe3UKZfLuP/++/GNb3wDGzduVLMdSafT1+RHAarGp2rMf+lLX7J9rsfu3bsxOTmJ6elplEolzM7OYmJiQi3WFNM0oes6AOD1118HP5DS7XZjamoKZ86cgWmaSi2CIAiCuPaQEdwlyuUyfv/3f18YAO3S398PALjllltEmsfjQTgcxvvvvw+fz4e9e/cuqD8p8zFzPB4PDh48CF3XMT4+jkKhYMtf7Bw5cgQbNmyomXcjjh8/ribNG88//zySyaT4EcUYa3msxWIRXq8XHo8Hbre7Iz02TRPRaBSpVAoHDx7EwMCALd/tdmNychLRaBTlctmWRxAEQRDXGjKCu8TXvvY1PP744/B6vWrWnHG73XjhhRcAYMH+GV6Gy+Djjz9WsxYt5XIZ+/btw7Zt29SsupimifHxcTV5XigWixgfH8f+/fsRj8c7MjLfeecdlMtlTExM4MSJE3jggQfUInUpFAoYHh7Gs88+2/Atucfjwdq1a3HkyBE1iyAIgiCuKWQEd4FcLgfLshAOh9WsrjEwMABd12FZlu0N68TEBLxeL1wuF7xeb9M/vReLReH7yZEDw+T2mrVVD26A3XDDDUDVQJJ9VYPBIIrFIqCMp1AowOv1iqC9RvXaJR6P2+TU7hv1kydPolKp1LxZLZfLiEajwo82nU4jl8shnU5jeHgYADA6OgqXyyX8uguFAsLhsJiX/GZUXgsul76+vqZrcezYMQDA9PQ0RkdHceutt7b1Jv6+++5DpVIR9aampnD+/Hmbf3Ejv/Rnn30Wmqbh8uXLNr9iJ2P8nnvuwaFDh9RkgiAIgri2sDa4cc197MY196nJPU2pVGKBQICVSiXGGGOBQIAFAgG1WEsAYIZhqMmCQCDAALBsNssYY8wwDObz+ZhlWYwxxhKJBAPAYrGYUvNfCIVCDADjS5/NZpmmaQwA8/l8ou1YLMYAiLbrIfdXKpVYJBJhAGwy0HWdhUIhxhhj+XyeAWCRSIQxZTypVIolk0mmaVrTeqw6dlVmTmmxWIwlk0nGlDHyuTnVUYlEIo7rmkgkxBgty2K6rgsZOrVrWRYLhUK2vjVNY5FIRJTn8svn84xJa8HnUI9SqcRSqZTQE13XhV42Ip/Ps1AoxDRNY7qu2/JCoRDTdb0lPZDHzHWRy0YmlUrZ5E8QBEEsLHrF3iMjeI7EYjFh9LBraASXSiVHw0g18Jzg7XCcjDWeJs/NCW5s6brOAAiDTja+NE2zjVOVER+ParA1q9do3DyNy8np4mXUOk6ofXMMw2ChUEiMPZlMNjSCuUHrdNWrw6o/CJz6r0erhnMymWQ+n4/l83lmGEZNHdmwrUc9Xamni/XKEwRBEAuDXrH3yB1iDvA/D6sBQPPFzMwMAGDVqlU4f/48AOCmm26yldm8eTMA4OLFi7b0+eQrX/kKZmZmwBjD7OwsDh48CLfbLfJnZ2cxNDQE0zTh9/trdjDgyHXQRr1GcDnJwWL8GhkZUYu3zZo1azA5OYlbb70V0WgUa9eubagPZ8+eFbtpqFcjQqFQW/Pfs2cPdF3Hhx9+qGYJcrkchoeH8cILL6C/vx8PPfQQNE0TbgzpdBr33HNPjQtIq1wPXSQIgiCIViEjeA4888wz2Lt3r81vMpPJIJPJwDUHn1onuN+xruvweDwi/cqVK7Zyy5Yts31eCJimiTvuuAPnzp3D4cOHEQgE1CKOdFrPiUb+rHNh48aNyGaz2LBhA8bHx6HretMtwE6dOqUmNWX58uVt79iwYcMG3HzzzWqy4MUXXwSkH3FutxtPPvkkKpUKtm/fjuPHj7f0Q2HVqlUAgAsXLtjSuS5y33CCIAiCWEiQETwHpqamat7mBQIBBAKBrr1phLT9GgAcPXoUkAyPl156yVaWGyKrV6+2pV8v+K4BBw4cwJ49e1p+q9hpPRUupy1bttiC4XK5XFv74Ho8Hscgr3g8jlWrVsE0TViWBZ/PJ9bICY/Hg0wmg2g0agvyaxZU+dprr2HDhg1qckOKxSKGhobUZMGlS5cAKZARAEZGRqDrOiYnJ2tkXi8o0ePxOM77woUL0HW9ph0OXxuCIAiCuB6QEXwNiMfj6Ovrq2tEoGr0AcAHH3wg0orFIkzTxK233orp6Wkkk0nx1s7j8SAWi2FyclIYc8ViEfv374dhGDWuBTLc6OH/8rfJly9fFmV4mvqmWYaPWa6nwrdJ++ijj4Dq4RHcrYMbpU7GZSv1Go2bp3E5VSoVbNq0SbyxHxwcFAaiWseJdevWYXp6Wk0Gqq4KxWIRHo8Huq5jxYoVgPSG9fLlyygUCjBNEzt37oSmaeKtMR+PvFsHqkYv15eJiQlYloWdO3faysjE43GxHvzEtq1btzbUg9tvvx0A8OUvf9m2O4XX64WmaXjkkUeEbhUKhYbbmr3wwgvIZDKiPD+J78CBA2pRnD59uuYvGgRBEARxzVGdhBvRK47Sc8EpgMowDKZpWk2AEIcHhjldPp+PxWKxunUNwxBBabqus0QioRaxofbFg5T4ZRiGCJCS01TUdnhbTvCyfOeEWCwmgt7kdtSdCRrVcxq3UxqH10VVpqlUijEpSMupjkypVGKaptXM0TAMls/nxVgD0k4hTAoOk3fsyOfzzOfzMTgEEvLx8F0Z+HibBafJcgyFQk3Ls+qcZLnous4Mw2ClUolZlsUikYjIk4P/6pFKpWy6yGWsEgqFGu5gQhAEQVxfesXec7FmETkSn/7cFwEAPzr3TTWLIJY88XgcH3zwAQ4ePKhmdY1cLofBwUEYhtE1d5qFRLFYhM/nw/T0NL0JJgiCWKD0ir1H7hAE0SIPPfQQzpw50zTwjahPLBbDs88+SwYwQRAEcd0hI5ggWsTtduPw4cOIRqPzZgi34p+8GCmXywiHw+jr62vr6GmCIAiCmC/ICCaINujv78f777+Pb3/722rWnMnlcti0aRMAYO/evV3dYu96c+zYMaxfv35eXUkIgiAIoh3IJ5ggCIIgCIIQ9Iq9R2+CCYIgCIIgiJ6DjGCCIAiCIAii5yAjeIFRKBQQDAbhcrng9XrnLQBrPjBNE36/f0n5snL44Q/BYFDNmjNjY2Pw+/1qso2lJlvTNBEMBuc0n2KxiLGxsZqDRq4V7c6hGzpULpcxMTEBr9c7b0eBE51TKBQQjUbVZKBDfe2GznSDVsbeiW52Y35O/TqloXrQUrODqzphbGxMHFZELDLUjYMb0SubJ3eCenhEJ4cB8AMZ8vk8K5VKLBAIMJ/PpxZbkGSzWXGwQr0DJxYzfG7qQSjdIBaLNVznpSbbbs0nFAqJ++1a08kcuqFD/OAdNDichrg+JJPJhgfbdKKv3dCZbtDK2DvRzW7Mz6lfpzRWPdBHa3BwVafk83mmVQ9yWir0ir1Hb4K7QLFYRCaTsaV96Utfsn1uhWPHjmHt2rXo7++H2+3G1NQUzpw5oxZbkAwMDOD1119Xk5cMs7Oz8Pl8anJX2LNnT806p9Np8Ybxesu22VvqdunWfEzTRCAQUJOvCZ3MYXZ2Vk1qm5GRETz55JNq8nVjbGys5bd+i5FyudzSW8pCoYCnnnoKU1NT6O/vV7OBqr7GYjE1uSHd0Jlu0MrYO9HNbszPqV+nNADYuHEjZmdnu75PeX9/P95880089dRT9EZ4kUFGcBd4/vnnkUwmwRgTV70HYSO+9a1vqUnEAsLtdqtJ88bx48fVpOtCsVjE9PS0mkwQAIDJyUk1aUlx8uRJNcmR0dFRPPPMM02fEcuXL1eTFg2LeezXgv7+fjzzzDMYHR1Vs4gFDBnBc6RYLGJ8fBz79+9HPB5HuVxWixBEW5imifHxcTX5mlMulxEKhdRkggAARKNRWJalJi8ZGvn3yuRyObzzzjsIh8NqFtFjhMNhZDKZJf3XkaUGGcFz5NixYwCA6elpjI6O4tZbb237zyG5XA4ulwuZTAaZTAYulwsul0vkcwd/HiwnB+LIAQuFQgFer7fhn68LhQLC4bDoIxqN2gz3eDyOvr4+uFwu9PX1YWJiom79vr4+jI2N2fI5pmmKdtQ2ZJqNPx6P2+aeTqdt9eV8v99fI/u5yM40TVG30RxQHQeXKV87pzQAIvAxHo/XBJyk02kMDw8D1bdLLper5oHaqmyhBFq6XC4Eg8GaoJB6MgwGg+ItMK8rt8WR2+a00m8rtKpvMo10WB6XaZq29srlMqLRKFzV+yKdTtfIvhmtzntiYkKMUb0H+Th4fjAYrNFrFT5nv9+PYrFYN1hPDhhKp9NC9/r6+mrurVwuB7/fL+Qoj3NsbEz8UBscHLTpg4wc+CTfT3Jbahl5nZvJotG8G9WV+5TnyXW4WCxi/fr1qFQq4plcT6avvvoq1q5dqyYDVX3gbdczktPptE3O9frhyHrqqj7XuJ7Kuievifws4mWb9dvK2FshnU6LdXdV196Jbt8TKk7Bcs30QK7LZS7rpxOBQACvvvqqmkwsVFQn4Ub0iqN0u5RKJZZKpURwnK7rrFQqqcWaEggEagIEDMNgPp9POPInEgkGKfBODlhIpVIsmUwyTdNsbXAsy2KhUEi0xYN7IpEIY9XADgAisIO3zctblsU0TWOJRIIxaSz8czabZQCYz+cTwQixWMzWhkqj8cdiMRFoUCqVWCQSsbXFA8osy2KlUonpum6b+1xkx//P55FIJJiu6zXrI8PnLwdHWJblmMaDqZwCTng7csBVJ7JljDFd11koFGKsGrwBQKw3a0GGXKdlfD6fLa1UKjGfz2eTTbN+neao0kzfmMP4mulwKBSq0VdOIpEQY7Ysi+m63jDAx2kOzeYNwKZXfLxymVAoJPIty2I+n4/pui7yDcNgkAJ+8vm8eOaUSiUWCoXqylUOGOLPAh6EK/eRrT4beB/8sxzAqY7DCVR1W9d1lkqlGKsGJ8lz5uPRdZ3l83kWiUSEDBvJotm8G9WV++T6wMcl36toIWhL13VHefNgKVnf1HsnlUqxWCwmvi+4jjcaQyQSsc1b13UxL/k5qQbocfmyFvptZexOqDrBn3/qPazOr9v3hFOaU7Bcq3rAg9RLpZLoH1XdVvU/FovZxrZY6RV7r7FGK/SKUOYCN0zkG6hVVCO4VCo5tqUag9wQ4A+0evCxOV1MMvQ42eqXPL/JI5GI7cHEjR/VqHAy3tQHhYzT+PncnS7DMMTDVTYAZSN2LrIrVXfpUOvqTYxgVjUQndZQTovFYrZxq4ZcIzk6pTWSrToPWceayZCXl8fWKE2eY6N+WZ35qDTTN+YwlmY6HAgEbH3K7RuGwUKhkNCFZDLZULZOc2g2b1UXmHRflqo/puGg8/Ic1C/3bDbLdF0X62hJP7KcUOvLaRyfz2eTDZOMEz4/p3accJozvw/lMrKBzSRjxOnKZrMN592sLu9THReU9XQqo6LW4QQCgRoZqvOWjSn5aqQzoVCoRm/lNvnzSy6TTCZtn5v128rYnVB1gv8QlJ8xaEHGc70n2klz6t9pjPJnvsuE03euuh6LlV6x98gdosvs2bMHuq7jww8/VLPa5vz58wCAm266yZa+efNmAMDFixdt6c2CMs6ePQvDMFD98WO7AGDbtm2YmZlBoeoLt2XLFlv9YrEITdPEZ7fbjTNnzmDbtm22cp0ij5/PXR0nYwwjIyNi7nKwRjgcFtHGc5Hd+fPnUalUaup6vV7bZycef/xxZDIZ8SfwY8eOwTAMW5plWV2PTq7H7OwshoaGxD7D8i4mzWQ4Fxr12yqd6FszHb799tsxOjoqXGP+8A//UOStWbMGk5OTuPXWWxGNRrF27VoMDAzY6jejk3l/4QtfAKp699577yEQCNToPGOs7lhWrVqF2dlZ6LqOcDiMixcvYmRkRC3WFtPT07jllltsaUNDQwDQlWcbvw9ldxP1+dVMFo3m3azutSCTydTIUP08PT2NbDZbM8apqSlbORnTNHHw4EGk02mEw+GaQCy3240nn3wS4+Pj4plz9OhRPPjgg6JMs35bGXsr9Pf3g1W/X7jbVSvM9Z6Yb5YtW4ZKpSK+Z4jFCxnB88CGDRtw8803q8kdc+XKFdvnZcuW2T63w6lTp9QkQbm6HdDo6CjWrVuHl19+WS1yzaPBm/lkNove7qbsWoEbCnxchUIBDz30EDRNw5EjR1AoFLB+/Xql1vxhmibuuOMOnDt3DocPH3bcUqyZDDuhlX5boV19a6bDe/bsQTKZhKZpNT78GzduRDabxYYNGzA+Pg5d12G2eVhNJ/PmOrpq1SoAwDvvvGPzh2yGx+PB9PQ0IpEITpw4gcHBwbp+l+1w+fJl22fVSJ0L/D7kc65HI1k0m3ejuguJ06dPq0kNKVTjF44fP47f+Z3fgWEYahHxzHn++edRLBYxMzNTYzC222+njI2NYcOGDVi+fDlOnDihZjsy13ui28RiMRw6dAjFYhHlchkvvvgiNE3D6tWr1aLEIoOM4HmgWCwKY2gu8AfASy+9ZEu/cOECALR9A3o8HmQyGUSjUVuwDg962L59O8rlMqamphAOh2sMxhUrVsCyrJqghbkETdSDz33Lli22gJ1cLoeJiQnccMMNQDVCXTaU0+k0TNOck+x42518SbjdboRCIRw+fBjpdBoPPvgg3G43wuEwJicn8dZbb3VFN1qhUChgeHgYBw4cwJ49e2q27Wsmw3rcfvvtapKNZv22Sif61kyHo9EowuEwzpw5g2w2C1TfeqH6pmrVqlUwTROWZcHn8+Ho0aO2+o3odN7Hjx+HruvweDy4+eabUalUagJ/4vF43UCgXHV3goMHD+L9999HJBKZ8+4iuq5jcnLSJnf+zFizZo1UsjPeeOMNMed6NJNFo3k3q9tNdF2v+cEAAJqmNXzpgGrd0dFRTExMCFnzIKx6rF+/HqFQCAcPHsTGjRvVbEB5GxyLxfDEE0/Y8pv128rYW2FiYgJ79+7F22+/jW3btjVcb5m53hPd5rHHHoPP58OGDRuwcuVKXLp0CW+++WbdH4a6rqtJxAKFjOA5It+IxWIR0WgUW7dutd0cPILZKUqcUy6XbReqRmssFsPk5KSIcC8Wi9i/fz8MwxB9tPoLeefOndA0Tbzl4tG6fdWdCS5duoRKpYJyuYxisYg33nhD1E2n09i1axcAYHx8HCtXrhSRybfddhsg/XqXvxB4mvpGVsZp/HzulUoFmzZtEmMdHBzE0NAQ+vv7EQqFUKlURHR6X18fduzYgaGhoTnJrr+/H4FAAPv27RPGoGmamJmZET8iGnHvvfdienoaL730knj78uCDD8KyLBw+fLjmi0D+EkL1IAZU5VgoFGCaZkey/fjjjwEAH330EVBdw5mZGfH/ZjIEgHvuuQeojo1HRHP3iUKhgHI14prLJhgMNu0XdXRFpZm+wUF2zXSY36PlchmrV69GX1+fzR0kFAqhWCzC4/FA13WsWLFC5Kmoc2hl3rquY2ZmRvzoME0TpmniG9/4BlA18H0+H6anp/H5z39e6P2hQ4eEUc37k9ed/5Bxu93o7++3uZGoONXnaVyOBw4cgGVZ2L17t3gmPf/88wgEAsLwuuuuu0Q7ExMTDZ9v77zzji0if3x8HAcOHADq3INoURb15t2srqo38v9lnQwEAmL+6u4JnA0bNuDs2bNqMp588klkMhnEq1tnFgoFvPbaa4DkWsVl8MgjjwgdX7lyJdatWwc46DcAVCoVsTVdoVAQxmqxWLQZhfxt8IkTJ/DAAw+IdLTQbytjd0LVLf75woULKJfLth/X831PtJLmJF8nPdi+fTt+53d+BzMzM8JtpN6P3FOnTmHDhg1qMrFQUZ2EG9ErjtLtwANzUI22ViNymRSVKgcHyPAAG/mSHfUNw2C6rjMoUaxM6b+ViNR8Pi+CInjwBHfu587+qO6gwCOEA4GAKJOVooQ1TRPBAuocDMMQAQJymkqz8cdiMTEmn88nIsxZNQBEzg8EAjXy71R2JSnKmgc78YCUeuvIqReUp/bPlDFAuh1537FYrGPZMql9vbrTAZcXH1szGXId0KXI8lJ1NwG5PA+k4WUa9es0n3rU0zdWR3bNdNgwDJZKpYROxKTjzQ3DEHPhc3MKfGF19J01mTeT9IqP0emZUarudMDblsuo656tBoilUimhM/JaqTjVV+XIg4ZSqZRN9jFpNwEmBSrK83MC1XtXlo1cnveraZrYFYLTSBbN5t2orjzfQCBQs56B6vOXBwO2ogtOGNKOBLFYjBnV4EsuY1btQ35GyTruNCa+hlp19wYeOCbrMicWi9UEuHEa9ctaHLuMk25Z1V1W+Pj5zgrq86Tb90SrafLnRnogp8mXk8w1TbN9Ty1WesXeczHutd4Cn/7cFwEAPzr3TTWLIAiCIGpwuVwIBAINg70WO8FgEFu3bm3oqnM98Pv9OHz4cN23lkRrRKPRui5G+XxeyNc0Tezfvx9nzpxRiy06esXeI3cIgiAIgpgDhmHgqaeeumY+qq3AXQ/IAJ4b5XIZt9xyS83OFIwxGIYhYivK5TKeeuopHD58WG2CWMCQEUwQBEHMC04+l0uR/v5+PPPMM1i/fv11NYQLhYLwmR0eHsYLL7ygFiHaZPv27Th16lTNTkXpdBo333wzPB4PCtWTIp955hn60bHIICOYIAiCmBdWrlwJVPelVY+iXWqEw2G8+eab+PrXv65mXTNuuOEGaJoGXdeRTCZrtkUj2mfXrl3weDzYsmWL+IHBdZm7v0xMTOCFF15YcO4wRHPIJ5ggCIIgCIIQ9Iq9R2+CCYIgCIIgiJ6DjGCCIAiCIAii5yAjmCAIgiAIgug5yAgmgGr09sTEBLxery0KdmxsDH6/31aWuLbkcjlEo1FbYFGxWERfX5/tSOlO6FY73aRcLi/aIKpevl/m8xkSDoeX/A4TBEFce8gIXoDkcjkRhSpffX19CIfD87IFz5EjR7Br1y5xHCexcNiyZUvdjdqXGoVCAXfccQd27NihZgmCwWDdKGz13uFHZtcjnU7byqvbIF0vTNOE3+8X4wqHw8jlcjBNc8GMUWU+nyH33nsvbr311nl59hEE0cOoR8g1oleO0Vso8GMiOfxIWE3Tao6V7Ab8+M16x2K2SiqVqjmCczGxEMfPj52dKz6fT01aMJRKpZqjb1UsyxJHltY7xpZJ947ucBS3DD/KVz2ut5u0o0/qUcR8jvl8Xsyp1bauB/P5DEkmk8zn8zVcd4IgukOv2Hv0JngBc9ttt9k+b9y4EQcPHkSlUsGzzz5ry+sGy5YtU5M64vjx42rSomIhjt/tdqtJbVMsFjE9Pa0mLxiOHDmCDRs2NNxs/vnnn0cgEAAAHDt2TM0W3HbbbQgEArAsS5ycpSK/VVTvtW7Sjj59+ctfxvT0NN58802Ew2Gx7v39/TBNE6FQSK2yoJjPZ0g4HIbf78eRI0fULIIgiI4gI3iRcdNNNwEALl26pGYtCEzTXNR/ul/s469HuVxe0AZUuVzGvn37sG3bNjVLUC6XceLECUxNTUHTNDz33HNqERs7duyApml46qmn1CygusH91q1b1eSu0o4+5XI5ZDIZRCKRuj8E9u7dqyYtORrJ7MEHH8S+ffvIP5ggiK5ARvAi48qVKwCAFStWANUvjGAwiHg8jrGxMbhcLhHkVCgUEA6HhV9hNBq1fXmUy2WR7/V6RducYrGIsbEx9PX12dLldvv6+jA2NgZU/SuHh4cBAKOjo131seTHgTYL4OLHV/I5Ob0FrCeXVscvy6VQKAjfzWAwKN4uqmW8Xq8tOCgej8Pr9YpxqvMyTVPkO/m1yusuI7fr9/vFeILBoHgLzMfaqB0e4MTHJ+fLcysWizYdKhaLolw6nYbX6xUy4HrixMmTJ1GpVOoaf6i++f3KV74CAIhEIrAsq6GP6LJly/Dkk0/Csqwa+RaLRZw4caKub3E9HUGDe069X1rVJ86rr74KANi8ebOaJfB4PBgZGVGT66KOta+vD319fbb1VMvI91kjOeA6PEMGBgZQqVRw8uRJkUYQBNExqn9EI3rFR2ShYBiGzfcxm80yTdOEz538ORAIMMuymM/nY4lEglmWxUKhELMsy1Y3EokwJvkeRiIRViqVRHnZn49/ltXEsiymaRpLJBKMMcYSiQQDID5ns9l58VvM5/MMAEulUmqWoFQq2eYoy4fPoZlcWhk/9yOV553P55mu68IHVZZdKpViyWSSaZrGGGMsFouxZDLJWHXMkUiEARBj4mX5OiQSCabruvAJlucljzMWizGfz8csyxL+tbxPJo2bU68dwzBEO0xa41gsxpgyN8MwhP7IcmSM2XzXDcNo6NMciUQa5jPGmK7rYkxcH+T+ZAzDYNlslpVKJYbq/SETi8XE2qnzb6Qjje45p/ulFX3i8PXh6z5XeN98rHwtYrEYA8CSyWTD+TSSA7uOzxDeJ0EQ80ev2HtkBC9guBEsG10+n89mCPIvDG6gcPgXndPFqm2rQUPJZNL2BcYcDKdIJGL7AuJfhK1+gc0nvG95/LFYzGYANZNLq+NX5cIc5MfLyIE83ChzurhBqVWDomRkI5g5jNOqBoxxg4VJxjTHacxqO3x8av+qoe7UViAQsI1RlgWrtlEPta5KMpmsqc8DyJwCpbgRzKSx88/8BwKvJ8+ftaEj6j3HHOSiyrcR3TaCWYP+1R9VTvNpJofr9QxppisEQcydXrH3yB1iETA1NYXqDxacOXMGGzduVItg+fLlts9nz56FYRiinnwBwKlTp+D1em11uL9xI4rFIjRNE5/dbjfOnDnT0JdTRd3Giv/Js176XFi+fDkymYz43Ewuc2FoaAgAcPr0aVu6HNR2/vx54KpFUHONjIzg/PnzqFQqNWuhrpXKxYsXAUUPwuEwZmdnpVLN4eNT++d/ouf9tILP58Pg4CD8fj9M08TBgwfVIi1z9OhRjI+P23Rjenq6pT+N79y5EwDwzDPPAJJbRb1gw1Z1RL3nWoW76/CLu6ZwF6drQSgUst0XcJhPMzlcz2cIQRBENyAjeAlz6tQpNakrTE5OqkkLgoGBAfh8Prz44osol8soFos4dOiQ2E2AM19y4UbVzTffrGbVMFfjvh7NDMJWUX07O4n6n5qagmEYqFQqGB4eht/v7yigifv9qoZYqVSCpmnYv3+/WsWGx+NBJBJBJpNBoVDAc889hwceeEAtZmO+dKQR9957LwDgjTfeULO6zvLly6Hruppcw3zJYaE+QwiC6C3ICF6ieDweZDIZRKNRW7ASDwRasWIF3nnnnbaNkhUrVsCyrJoAmXoBRk4MDAzUGDQDAwN109vh8OHDuHTpEm699Vbouo4NGzbglVdeEfnN5DIXeDDR2rVr1SzBqlWrgOoBGHKwVi6Xw8TEBG644QbA4W1yM3i9aDRqM7DT6bRjcGA9+PheeuklW/qFCxcAAKtXr7alN2L37t0YGRnBzMwMEokEpqenxZtmFY/HU1cX6+3i4Ha7EQ6HMT09bVtLJ/jb4Pvvvx8bNmyAx+NRiwjmU0eg/GWHMYapqSmg2r7P58P4+HjdgL9CoVATxNgJr732GjZs2KAm22gmh+v5DGm0fgRBEC2j+kc0old8RBYKPKjEyeeRwzenV/3nePCJ6svHffG43x0PfCmVSrYDBrhPn8/ns42BByTJl67rtv5R9S/M5/M1vqWd0mpgnOzr6UQzubAWx8/9HLkfIw8okv0quexU6vlayv62sl9wMplkuq7bxsnXXe5PDkICwDRNs8mD+5iXSiVRz6kdPj51bvIaq3rB03zSYRxyG3z96h2EwX1JVXjQVL015eOXA/lYVRbq2vE1k8txH+iQdFhGMx2pd8+xOnLh8m2kT5x8Ps+06oE4SemwjFKpxJLJpG2crcDvc1k+iUSCaZomPtebTzM5XK9nCBx81gmC6C69Yu/Vfus0oFeEcr3hXy7y5RQIIpdTv0RY9cuGfwHxqG75yzmVSgnjKhAIsFQqxXw+n/iC4UYDvzjZbNbWrtovD0RSA23mQitGsJPc+BjlL81mcmll/Fw23GDUNM1WXpadGjzEqoYmNzB8SrBjSdoxgo89FAqxSCTCLMuqmacc1Ca3G5B2BGCSgaVXT2Wr1w6Tgp74+Lkxy+rohfyZpxmGwQzDsBl19ShVAwK54cQko51fcp5TPgD2x3/8xzVpnGw2a7uP1HnIfdTTkUb3nNoepxV9kimVSjb5o6oj8hq0Ch9vKBQS7fl8PqEXjebDGsiBc62fIfw5UO9HEUEQ3aFX7D0Xu/oF1hKf/twXAQA/OvdNNYsgrjvlctm2H66MpmltB4k1IhgMIpPJoI3bh2hCPB7HBx98MKcAOsJOLpfD4OAgDMNoa3/hhUo0GoWmadizZ4+aRRBEF+kVe498goklw4ULF/D444/b/C1ZNYAqEomoxYkFxkMPPYQzZ8605cNM9A6FQgEnTpzAY489pmYRBEF0BBnBxJJh/fr1OHfunC2oqFwu4+TJk/jCF75gKztXeEBPu0FBRH3cbjcOHz6MaDRKhnCX4Lt8XL58Wc1aVKTTadx///34xje+UXdrO4IgiHYhI5hYMrz88suwLAuf//znxXGs27dvx+rVq9veZaIRssvFypUr1WxiDvT39+P999/Ht7/9bTWLkFD3Gq53bdq0CQCwd+/eruwqcb04cOAATpw40fBYbYIgiHYhn2CCIAiCIAhC0Cv2Hr0JJgiCIAiCIHoOMoIJgiAIgiCInoOMYIIgCIIgCKLnICOYWPTkcjlEo1EEg0GRViwW0dfXZzuauBO61Q5BEARBEAsLMoKJRc+WLVswPj6uJhMEQRAEQdSFjOAlSDqdXtTbIbXL7OwsfD6fLc3j8WB2dhYbN260pTfD7/fbPnfaDkEQBEEQCxsygpcgx48fV5OWPN3YQL9YLDoeuUwQBEEQxNKDjOAlhmma5BrQAeVyGaFQSE0mCIIgCGKJQkbwAsU0TQSDQcTjcYyNjcHlcongrEKhgHA4LE6FikajKJfLSKfTGB4eBgCMjo7C5XLh4MGDohx3kYjH4yItl8s17E9On5iYgNfrtbXlRLFYxNjYGPr6+lAoFOD3++FyuRAMBsWRxmoZr9drc0WIx+OiL6/XWxOYZpqmyJ+YmLDl8Xw+bhm5Xb/fL8YjnwLHx9qoHVkWXq/Xli/PrVgsirXyer0oFouiXDqdhtfrFTIYGxsTeQRBEARBzDOsDW5ccx+7cc19ajLRZbLZLNM0jQFggUCAWZbFfD4fSyQSzLIsFgqFmGVZtrKRSER8BsAMwxDtpVKpmrRkMskAsGw2W7e/nTt3inSfz8ey2SxjjLFYLMYAiDGoBAIBBoABYIlEgjHGWD6fZ7quM13XGWOMhUIhUSaVSrFkMsk0TWOs2n4ymWSMMVYqlVgkErH1x8vy8SQSCabrOgsEAowp8pPnHIvFmM/nY5ZlsVKpxHRdF30yadyceu0YhiHaYdX+AbBYLMaYMjfDMFipVGKWZdnWiTHGNE1j+XxetMnHTxAEQRDXk16x98gIXqBwY5YbVhxugDpdcj3ZaGuUxg3Jev21UtcJ1aBkiuEtlymVSqJMqVSqmRe/uEGpaZowkjmyEcwcxm1ZFoNiuMuGN6szZrUdPj61f9VQd2orEAjYxijLglXbIAiCIIjrTa/Ye+QOscBZvny57fPZs2dhGAaqP2BsVzdQ++smQ0NDAIDTp0/b0uWgtvPnzwNXrceaa2RkBOfPn0elUsFNN90ktQB4vV7bZ5WLFy8CyvzC4TBmZ2elUs3h41P737x5MyD10wo+nw+Dg4Pw+/0wTRMHDx5UixAEQRAEMU+QEbwIOXXqlJq0KODG7s0336xm1cB9lbvNyZMn1aSOuHLliu3zsmXLbJ9bYWpqCoZhoFKpYHh4GH6/H+VyWS1GEARBEMQ8QEbwIsPj8SCTySAajdqCrMLhsK2czOrVq9Wk6wIPblu7dq2aJVi1ahVQPQBDDobL5XKYmJjADTfcADi8TW4GrxeNRm0GNg/+axU+vpdeesmWfuHCBaBNWe/evRsjIyOYmZlBIpHA9PS0eNNMEARBEMT8QkbwAkV908jZuXMnNE3D+Pg4dF0Xuzz09fUBAAYGBgAAly9fRqFQgGma4g3su+++C1QNyldffRUAsHXrVpimWbc/nn758uWatHp1ZPjODcViEU8//TRisRg8Hg9Q3ZZMxePxIBaLoVKpYNOmTWJ+g4ODGBoaQn9/PwKBAPbt2yeMV9M0MTMzI34cwGHc/f39CIVCqFQqGBwcFDLbsWOHcNO45557gOq4+E4Najt8fJOTk7a57d+/H4ZhCFnzuclzLJfLts/j4+OijTvvvBOQjHWCIAiCIOYZ1Um4Eb3iKH294cFYAJiu67agNFbdacHn8zEAYscBObiMB2nJQW48CEzTNGYYBstms0zXdZZMJtnx48cd+5PHwYPDDMOoSXOCB4bxQD5N02zjkXeQ4DtGyMRiMdvOFKlUSuTJO0bwILlQKMQikQizLMtx3Lye3G4gEBC7M7CqXDVNY7qus3w+X7cdVt3NQdd1MX6+CwZT5obqLSZ/5mlcnnxd1GA7giAIgrge9Iq952JtRFR9+nNfBAD86Nw31SyCsBEMBpHJZLoWsEcQBEEQxLWhV+w9cocgCIIgCIIgeg4ygol5wcknliAIgiAIYqFARjDRdeQjiFeuXKlmEwRBEARBXHfICCa6ztTUVNcP8SAIgiAIgugmZAQTBEEQBEEQPQcZwQRBEARBEETPQUbwAqNQKCAYDMLlcsHr9bZ1mlkvEI/H0dfXB5fLhXA4vCgD73K5HKLRKILBoJo1r4yNjcHv96vJNkzThN/vRzweV7MWJaZpIhgMzmk+xWIRY2Nj4kCaa027c+jWeNvt93rh9/vF4Tbo4P4ql8uYmJiA1+tteFx7u+0SBLHwISO4ixSLRUSjUWGktUu5XMb69ethGAZKpRK8Xi/279+vFutZ4vE43n33XczOziKZTOLEiRN4++231WILni1btmB8fFxNvu7wL3ke1LjY4fPJZDJqVlvEYjHs3bsXlUpFzZp3OplDN8bbSb8LhXbvryNHjmDXrl2wLEvNstFuuwRBLHzICO4SpmlC13UAwOuvv95RQNixY8ewdu1a9Pf3w+12Y2pqCmfOnFGLLSqavXlsh3379uHRRx8FAITDYczOzmLjxo1qsQXP7OwsfD6fmjzv7Nmzp0af0um0eNM3MDCA119/3ZZ/LemmrqCL8zFNE4FAQE2+JnQyh26Mt5N+rxdnzpzBnj17xOfZ2VlbfjNGRkbw5JNPqsk1tNsuQRALHzKCu4BpmohGo0ilUjh48CAGBgbUIi3xrW99S01a1BSLxa69VczlcnN6s7XQcLvdatJ14fjx42rSdaGbukIQBEEQrUBG8BwpFAoYHh7Gs88+uyjfSs4X5XIZoVBITSYWEKZpLog/75KuEARBENcDMoLnyLPPPgtN03D58mXhCxyNRtsK2MrlcnC5XMhkMshkMnC5XDafYh60wYPl5EAVOQimUCjA6/U2/LOyHFjW19eHiYkJtYiNdDoNr9cr2pcDUFBtTx5bOp0GlAMzXC5Xw2CSYrGIcDgs5h0Oh1EoFER+MBjE4OAgAGBwcBAul8sxWKeZLOqN1Snf7/fbxoA5roNpmqKuk8ybyZkTj8eFnLiOOKWhKjcuKzVYKp1OY3h4GAAwOjoKl8tVExRkmqbQFacxy8gBna7qeheLRVuZevJ10hW5LY7cNqeVfluhUCgIHezr66srf5lG91K5XEY0GoWr+jxIp9NCvo3yWqXVeedyOSHzYDBYo9PN7gmVVvWUU09G8thdku7JugxFVnyMsqxaDVZr1g7nwoULYmytyANVmfj9fjFHp2cTQRALFNYGN665j9245j41uacBwAKBAMvn84wxxhKJBAPAQqGQWrQpgUCABQIBW5phGMzn8zHLshiT2o/FYowxxkKhEAPAALBUKsWSySTTNM3WBieZTDIAYqy8Lm/bCU3TRHnDMGzji8ViLJlMMsYYK5VKLBKJ2NoLBAKsmYqVSiWmaRpLJBKMMcYsy2I+n49pmmYbVzabZQBYNpuVattpJItmY43FYkLOpVKJ6bpuk+Nc1oH/n489kUgwXddtsmwkZxUuCz4fVpWbU5phGIwp4+PwdngZOc3n84nxxmKxpnqi67rQ+Xw+zwCwSCQi8pvJ10lXfD6fLa1UKjGfz2eTTbN+neaoYlmWTQf52vLPzGF8ze6lRCIhxmVZFtN13bb+9fKccJpDs3nz8fI5ZLNZpus603VdlGl2Tzj1246eNpNRKpViqN4vMrFYjJVKJcYYY5FIhOm6zkqlktAbeQ6apjFUn8EyalqzdgzDEDIslUrMsiwhQ1nv1XZTqZRtvFx35PuQIBYjvWLvNbZQFHpFKK1SzzBTv0xaRTWCS6WS4wNVbZ8/rPmDuB7c+OLUG7+Mms+/aPnYnC7+pakaDk7EYjHbmJhk0DkZM43GyurIotlYeX/yeslG7FzWoVQ18tW6umIEq3OT5+6EagzyMcppsVjMNid1PZyMnEZpjWSvzlHW5Wby5eXlsTVKk+fYqF9WZz4qkUjEJu9S1dhuZAQ3u5cMw2ChUEjoQjKZbCnPCac5NJu3Ol6mGJ3N7glWp195jqyJnjaTEavqsdx+qVSyzSsUCtn64MaqDFowgpu1wz/LY+My4j90mUO7/IeaeqnjIYjFRq/Ye+QOMQ9s3rwZAHDx4kU1qy3Onz8PALjpppts6fXabxZstW3bNszMzKBQKCAajWLLli1qkRp8Ph8GBwfh9/thmiYOHjwISGOTj0fm18jIiNJKfc6ePQuv12tL83g88Pl8jn/ebRVZFs3GyuW4fPlyUYfvPiHX72Qdzp8/j0qlUlNXnXM9Odfj8ccfRyaTETI6duwYDMOwpVmWBY/Ho9ScH2ZnZzE0NCT2GZa31mom37nQqN9WKRaL0DRNfHa73Thz5gy2bdtmKyfT7F5as2YNJicnceuttyIajWLt2rUiYLZRXqt0Mm8es/Dee+81vSfq0Y6eNpMRqnq8b98+8fnkyZMYGhoSn3kf6XQa4XAYo6OjIq8dOmnH7XYjEAjg7NmzapZgenoa2Wy2RoZTU1Nq0Rq3pXoXQRDXDjKC58CqVauAqh+ZzLJlywAAN9xwgy29U65cuWL7zNtvl3K5jGAwiNHRUaxbtw4vv/yyWqSGqakpGIaBSqWC4eFh+P1+m7+zk19duzj5Tzcz6Duh2VhPnjypJtno1jo40UzOKtxQ4GMuFAp46KGHoGkajhw5gkKhgPXr1yu15g/TNHHHHXfg3LlzOHz4sOMWXc3k2wmt9NsKk5OTalJDmt1LGzduRDabxYYNGzA+Pg5d12FWD75plNcqncyb69PNN98s0prdEyrt6GkzGaGqx5VKRcz/3Llztnuf+9cfP34cv/M7vwPDMKTardNpO+VyuekPydOnT6tJjoyMjNQYy04XQRDXjraM4J/92f8VAPA//+c/qlk9CX9jefToUVv6hQsXoOs6+vv7bentwo3sl156yZbOje7Vq1fb0puxfft2lMtlTE1NIRwOt2TE7d69GyMjI5iZmUEikcD09DTOnz8vxrZlyxZb8Egul2saRCXj8XgwPT1dE7AzMzOD22+/3ZbWKc3Gyn+sRKNRm1GQTqdhmuac1oG33eyLsp6c6+F2uxEKhXD48GGk02k8+OCDcLvdCIfDmJycxFtvvWV7ozafFKo7pBw4cAB79uyp0ftm8q1Hs/Vv1m+rrFixApZl1QS0hsNhWzmZZvdSPB7HqlWrYJomLMuyPSca5bVCp/PmP0LWrl3b9J6oRzt62kxGqOpxJBLBU089hXQ6jbvvvtuWv379eoRCIRw8eHBOu+900g7ftm/dunVqlkDXdYyOjmJiYkLoDg/CI4jFCrfxuM23lGnLCP7Uz38CAPCTv/t7NatneeGFF5DJZMQXB49WPnDggCjDI6Qb/Xm/XC7bLlQNxFgshsnJSdF+sVjE/v37YRiGeGNS702MyqVLl1CpVFAul1EsFvHGG2+IvHpR0OPj46LvO++8E6gaNXxslUoFmzZtEn/KGxwcFMbXPffcA1THVy+KfOfOndA0DQ8//LCQz8TEBGZnZ/HYY4+JctzgVN+6qzjJotlY+/v7EQqFUKlUxO4TfX192LFjB4aGhua0Dv39/QgEAti3b58w+EzTxMzMDDKZjPiyrCfnRtx7772Ynp7GSy+9JP6c/uCDD8KyLBw+fLjmDZb8JY3qgQgAcPnyZRQKBZimKd52X758WdTjaeqbcM7HH38MAPjoo4+Aqi7NzMyI/zeTL+roCnefKBQKKFePtuVyCwaDTfuFNGZ5Piq7du0CqmuwcuVKuKo7A9x2222ijCq7Vu6lUCiEYrEIj8cDXdexYsUKUaZRnoo6h1bmzdee6xx3SUgmk/B4PE3vCTj0izb1tBUZQdLZp59+usZArVQq4iS3QqGAU6dOAdX7j+sFlHvPKa1ZO/zt+BtvvIFy9RkcCoUQiUTEjyGndvlz/pFHHhG6s3LlyoaGM0EsdLiNx22+JY3qJNwI34Zt7MY197EfXPiRmtXTpFIppus6A8B0Xa+JdjYMo2a3AxkeMFIvsMIwDFv7TgE7PK8RqVRKRFPHYjGWz+eZpmksEAjUDaozDEOMXw3GYdXgK96mz+ezzZ23r+u6iBB3Ip/P23YvCIVCjgFd8uUUSNRMFo3GWiqVbPkBaccPTqfrUJIi77kMeaAOn2czOTtRL2BPHRtzkCGHjysWi9XoIR+TmuYEb1+v7nTAZcnH1ky+TrpSKpVEu7x8IBBgkUhElGnUr9N86pHNZkWQk6ZptrJOsmt2LxmGIcbLx8/vsUZ5KvXm0GjeHFlfVX3n1Lsn6vXbjp42k5GMruu2ADQO1z+tunsHD+7jZeUxBqrPTKe0Zu2w6ni5Dui6XqMvTu2yamCj/FxQ6xHEYuMHF37EblxzH/Nt2KZmLTlcrA0npPVf/H2cf/8HePObL2D1rZ9VswmCIAiCIIhFzPn3f4D1X/x9rL71s3jzmy+o2UsKcocgCIIgCIIggB5zh2jTCP45AMClyz9RswiCIAiCIIhFDrfxuM23lGnLCNZv+UUAwMwHf6NmEQRBEARBEIsc6wdX93bnNt9Spj0j+LNXt9UpfvBDNYsgCIIgCIJY5Fh/XTWCqzbfUqYtI9hzy1WB0JtggiAIgiCIpQe38bjNt5Rpywj2Vl+NF6uvygmCIAiCIIilA7fxuM23lGnLCP7Fz6zEJz/xc/jbH1dQuXR1w3aCIAiCIAhi8VO59DH+9scVfPITP4df/MxKNXvJ0ZYRDAC3/5v/DQDw3TPvqVnEEmNsbAx+v198LhaLGBsbQ19fn62cDD/Zy+v12o7I5TQ7KrfbNBsPQRAEQRBX4bYdt/WWOm0bwb/m+1UAwF9Of0/NIlogl8uJI0rVy+/3Ix6Pq1Ua0qi9vr4+hMPhrhl/sVgMe/fuRaVSUbMER44cwa5du8QRpSrHjx/HuXPnasaqzlvN73QOzcZzvQkGg+JYVhV1bflxtfVIp9O28gcPHqyRo9MFh77ky0kvncoHg0GgOic1T7142VYwTRN+v1/U5TptmmbHekEQBEHUwm07bustedQj5Jrxl2e/x25ccx/79fv+k5pFtAE/JpiTz+fFkZ1Ox4c2Q22PKcc5Ox2Z2gn8uNZG8CNJnY42jkQi4v8+n4/pul5zhCqHHy88VxqNpx6pVGrejz+1LEscw1pPBkxaW/U4ZhW+NqrMnHSDVftX21TLNtNLfuyyfMy1z+ezyToQCNiOmc1mszX9OlEqlZjP5xPH83IZycdsz/caEQRB9BK/ft9/YjeuuY/95dnvqVlLkrbfBN9526/gk5/4OZx//wf40Y9n1WyiRW677Tbb5/7+fkxNTQEA9u7da8trBbU9ANi4cSNOnDgBADhw4ICaPW8sW7ZMTQKqbyrXrVsnPrvdbni9Xrjdbls5zk033eQ4r3apN55GHD9+XE3qOs8//zwCgQAA4NixY2q24LbbbkMgEIBlWXVdSQqFgvi/KjP1M8fj8eArX/mKLU0t20wvb7nlFqDaFsfv92NgYEAqZWdgYAA+n09NruHLX/4ypqen8eabbyIcDgs96e/vh2maCIVCahWCIAiiQ37041mcf/8H+OQnfg533vYravaSpG0jGAB+zXdVOLm3z6lZxByoZwwuFY4fP46hoSE1ecFhmibGx8fV5K5SLpdx4sQJTE1NQdM0PPfcc2oRGzt27ICmaXjqqafULADAxMQEtm7dqiY3JB6PY2RkRE2uoV29PHjwoJpUQz1jnpPL5ZDJZBCJRNDf369mA3WMcoIgCKIzuE3HbbxeoCMjeGjd1bc4336D/PG6SbFYBADouq5moVAoIBwOC7/IaDSKcrmsFrNRKBTE27IdO3YAii8n9/OMx+MijftYthIExymXy2JsXq8XV65cUYsI2jWoZEzTRDAYRDweF2Pr6+ur8VdtZTzxeBx9fX1wVX2nub9tOp3G8PAwAGB0dNQmk07WoB7Hjh0Tb2EjkQgsy7K9zVVZtmwZnnzySViWhXQ6bcsrFos4ceJEXd9iJwqFAk6dOqUmO9JIL+eLV199FQCwefNmNUvg8XhaMuIJgiCI5nCbjtt4vUBHRvBvBu8GAPz5X5zFD/+2MyOA+BfK5TJM08SGDRugaRq+8Y1v2PKLxSKeffZZ7N27F4wxZLNZmKaJ3bt328pBCSj7/Oc/D8uykEwmsXHjRqD6p+hUKmWrMzIygmQyaUtrJQgO1bEHg0H09fWhVCrhxIkTeOmll9RiNa4Q7ZLL5TA8PIxMJoNTp07hS1/6EmZnZxGJRDA6OireLLYyHtM0MTo6ijfffBOMMWzYsAGPPPIIisUiNm7ciGw2CwAwDAOMMQwMDLS1Bq3w3HPP4YEHHgAAfOlLXwKqb3Mb8dBDDwEOri1HjhzBE088YUtzQtWNZjTTy/mEG96duLIQBEEQ7fHDvy3jz//iLCDZeL1AR0awtuIG/NsNdwEAvkVvg+eEy+XCypUrMTw8jK985SuYnp6u+fPvkSNHMDk5CV3X4XK5MDg4iEql4vgne8aYuFKpFHRdx/DwMMbGxkQZJ8Pipptusn02TVP4qzbiyJEjqFQqOHjwINxuNzweD+6991612JxdIQYGBoRxes899wgZ7dmzB7qu4+jRo0CL47l8+TJ0XRdtPProowCAixfrHwLTzho0gxuW3I+2v78fPp8Ppmk2fLPsdrsRiUSQyWTE2+lyuYzJyUlhUDdC1o18Pq9m22hFLwmCIIilAbfl/u2Gu6CtuEHNXrJ0ZAQDwL1fuBr48p2p76pZRBtwoyQQCOC1116zBRhxzp49K95KqlcjNm7ciKmpKei6jr179zb8c3unnDp1Cl6v15amGtScubhCNCIUCiGTyQAtjmfbtm2YmZlBoVBANBrFli1bbPlOdLoGThw9ehTj4+O2N7PT09OoVCo4efKkWtzGzp07AQDPPPMMILlVtCvb/v5+3HPPPWqyoBW97Abqdmp867QVK1aoRQmCIIh5gttyvfQWGHMxgjf9xp3oW7EMhe/N4K2z31eziTZ55ZVXYFkWotGomgVUjbtOcLvdwveUG4rXmnquELfffnvDMV24cEFNcmT58uVt+atyl4nR0VGsW7cOL7/8slrEkU7XQIb/EFEN6VKpBE3TsH//frWKDY/HI94GFwoFm1tFu7TiT9tML+cL/vb+jTfeULMIgiCILvLW2e+j8L0Z9GnLsHno19TsJU3HRjAA/Pa/u/rn7cOv2H1MifZxu914/fXXMT4+XhM57/F4kMlkEI1Gha8kgJYDobjxtmbNGgDA6tWrlRKds2LFCrzzzjsN/4xfzxXic5/7HFD193Xi6NGjLblkvPbaa9iwYQPQ4ni2b9+OcrmMqakphMNhR/cQlbmuAafeLg5utxvhcBjT09O29p3gb4Pvv/9+m1tFpzQycBvpZTeYmpqy/Rjg27GFw2H4fD6Mj4/X/QtGoVCoCYokCIIg2oPbcL/9QO339JJH3Ti4HX5cvsRW9d/PblxzH/urc++r2UQD+Gb/6iEJhmGIgwn4AQSWZTFN08TBCvySD57g7cmHFsiHCsiHFbCrf8MXhypks1lx6IGu6yyZTDJWPfRAHaOals1mRVuWZbFSqWQ73IG3XQ9+YIZ8mEcqlWKBQKCmHu/L5/OJeSYSCaZpmvjcyngCgYA4pMOyLBaLxRiqh2nwcaC6Bvl8niWTyZbWoBmJRKJGnjL8UA95fqy6tnxNOPxgDLlcqVQSc5epp2uWZbFQKGQ7BKNeWSe9ZHUOy5AplUpM13Xm8/nUrKbk83mmaVrNYRmlUqlrB6kQBEH0Mn917n1245r72Kr++9mPy5fU7CXPnIxgxhh7Ov5f2Y1r7mPR0RfULMIBbqTJl2qgcgMH0olY8sldmqaxSCTCSqWSY3vy5fP5mGEYNUZNMpkUBoZhGOIUL25syGNA9beSUxpTTqYLBAIslUoxn8/HkskkS6VSNQacTKlUYpFIRNSXx6wiG7i8vM/nY/l83lau0Xh4PjdouaGraRoLBAJCTty4kw3EemvQCtyI5FdWOb1OzQfA/viP/7gmjcONeY66NvXqO12WZTnqUSO95PJpVN5pTk7r2ohSqcQMw6jRj0QioRYlCIIg2iQ6+gK7cc197On4f1WzegIX6ySyR+JvfljC7Ruu+py++c0XsPrWz6pFCKIr5HI5DA4OwjCMlvxZCYIgCIJw5vz7P8D6L/4+AODsiUP4xc+sVIsseebkEwwAv/iZldgauhrR/fzBSTWbIAiCIAiCWGBwm21rKNiTBjC6YQQDwOORL+GGT30SqT9/i06RI+YNfvLb5cuX1SyCIAiCIFrk22/kkPrzt3DDpz6JxyNXD2zqRbpiBP+CewWe/L0HAQD7/uhV/PM/z8nDgiBqyOVy2LRpEwBg7969C2pXAHWvW6eL739LEARBENeTf/5nhn1/dPVo+id/70H8grt392Wfs0+wzL1bn8JbZ7+P33voixj7T7+lZhMEQRAEQRDXkT3/18v4oyPfxJ23/wq+ffTqwUu9SlfeBHP42+A/OvJN5N4+p2YTBEEQBEEQ14nc2+fwR0e+CUg2Wy/TVSP4ztt/Bdv/w30AgCe+msCVn/xULUIQBEEQBEFcY6785Kd44qsJAMD2/3Af7rz9V9QiPUdXjWAA2P37v42BOz6Hv/7whxipCpsgCIIgCIK4fox8NYG//vCHGLjjc9j9+7+tZvckXTeCAeC5P3gEyz71SXz7jRwOHv2Omk0QBEEQBEFcIw4e/Q6+/UYOyz71STz3B4+o2T3LvBjBv3zzZ/Dc0xEAwFf3/ylSf/6WWoQgCIIgCIKYZ1J//ha+uv9PAQDPPR3BL9/8GbVIzzIvRjAA/GbwbvzH370fAPDwY8/hu++8pxYhCIIgCIIg5onvvvMeHn7sOQDAf/zd+/GbwbvVIj3NvBnBALBrx5fx778UAGMMDz0Wx3v/46/VIgRBEARBEESXee9//DUeeiwOxhj+/ZcC2LXjy2qRnmdejWAAMHZvw28G78alyz/Bw489h+m/+n/UIgRBEARBEESXmP6r/wcPP/YcLl3+CX4zeDeM3dvUIkS3D8toRHjbV/HfTudxyy99Goe/9gTW/OtfVosQBEEQBEEQc+C9//HXePix5/DBRz/Cr9/1eZgTf6AWIarM+5tgzuGvPYFfv+vz+OCjH+GBh58mH2GCIAiCIIgu8t133sMDDz8tDODDX3tCLUJIXLM3wZxtT+zHd6a+C5fLhcNfewKbfuNOtQhBEATx/2/vbmOqLOM4jn/DcDwcFVGmwhok2CaoNcHk4cQraBA0sFZugFsLWVnDjYhmL6xlL2RQa8nyYWRzE9hsmbAOwYJX7PCU0CoEt8SCGqADAeEAThJ6AQc9d20BFhzg99nOzva////3v+u+d12XiMgclFc3cPCtfCYnJ0mKi+JMfraxRQwW7E2w3Zn87Pub5bLydI6wiIiIyEM4ea6M9Kz7m+AUgGdnwd8E2x0/UcynhRcBSI43kzd9wYaIiIiI/Lsh2yjvHDtNaYUVpo9B0ykQs7doIRigrLKWnA9OMWQbnbpg473XMe/daWwTERERkQdYG1vImb4Kea3Jg/z3D+kc4Dla1BAM8NvvPeQcO421sQWAzPQXePdwKi4ujxhbRURERFa0iYlJjp8opuDs1wA8E76LvKOv6Sa4eVj0EGz34Sfn+eyLSwAEPLaZI5kpJMebjW0iIiIiK1JphZXcghI6/rgBwJuv7uNo1gFjm8yS04RggIbmNnILSmhobgMgISact9/Yz/Zt/sZWERERkRXh6rVOPjp5gfLqBgDCdwdz5HAK4aHBxlaZA6cKwXZnS74lt6CEYdsoAC8mRJORlshTO4KMrSIiIiLL0o9X2ikssnCxvAaANSYPjmSmkJ7ynLFV5sEpQzBA761BPj71JecuVM7UEmMjOJiaoJWPiIiILFsNzW18XlyOpap+pvbK/jiyD72MzwYvh16ZP6cNwXZdPX0UFlkoLLZw794EAE+GBJEUF8W+eDNbNm0wjoiIiIgsKT03b3GpwkpZZS0/tbYDsGqVCxmpiWSkJeK3ZaNxRB6S04dgu77+2xQWWTj/VRX9A0Mz9ZjoUJLjzZj37mSzj7fDjIiIiIizutHbj7WxhdIKK9U1zTN1b6+1HHgploy0RDZ6r3OYkf/OkgnBD7JU1VNWWcs339U51Ldv8ydyTwgRYSFE7dnBeq81Ds9FREREFsvA4DC1l69Q39RK3eVWrl7rdHj+/LORJMebSYgJd6jL/2NJhmC7gcFhyiprqappor6pjdGxOw7PN/msZ6u/L0EBfgQ+7kugvy9e60yYPN3x9HDD5OGOydOd1atdHeZEREREZuvu3XFsI2PYRscYGb2DbWSMwds22ju6+LWjZ+q/s5ubvQMOcx7ubkSEBRMbHUZSXJRe3i2wJR2CjRp+aKPu+1bqm1pp/vmXv4ViERERkcXi4e5G6K4niAgLIfLpEMJ3a6P/YlpWIdioq6dvehXWzfXObq53dGEbmVqh2VdstpExxsf/NI6KiIiIzIqr66OYPN1nvjBP/dwIDPAj0N+XrQFTX6W1uc25LOsQLCIiIiLyT1yMBRERERGR5U4hWERERERWHIVgEREREVlx/gJcIzcLMWoHZAAAAABJRU5ErkJggg==\"\u003e\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cstrong\u003eAlgorithm 4.2-\u0026nbsp;\u003c/strong\u003ePart I \u003cstrong\u003eTrain AMERIT-C - building classifier pool- I\u003c/strong\u003e prequential evaluation of classifiers\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cimg src=\"data:image/png;base64,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\"\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAlgorithm 4.2- Part II:\u0026nbsp;\u003c/strong\u003eselecting best performing classifiers from Part I output, keeping classifier pool ready\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;Train AMERIT-C - building classifier pool- II\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cimg src=\"data:image/png;base64,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\"\u003e\u003c/strong\u003e\u003cbr\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAlgorithm 4.3 Part- I : AMERIT-C -Update/Rebuild\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cimg src=\"data:image/png;base64,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\"\u003e\u003c/strong\u003e\u003cbr\u003e\u003c/p\u003e"},{"header":"5. Experimental works","content":"\u003cp\u003eAll the experiments mentioned herein are carried out on the platform scikit multiflow [\u003cspan class=\"CitationRef\"\u003e17\u003c/span\u003e], and Anaconda Jupiter platforms as mentioned in Table 1. Following are the minimum requirements to carry out the experiments.\u003c/p\u003e\n\u003ctable id=\"Tab1\" border=\"1\"\u003e\n \u003ccaption language=\"En\"\u003e\n \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e\n \u003cdiv class=\"CaptionContent\"\u003e\n \u003cp\u003eExperimental Setup\u003c/p\u003e\n \u003c/div\u003e\n \u003c/caption\u003e\n \u003cthead\u003e\n \u003ctr\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eSr. No.\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eParameters\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eValues\u003c/p\u003e\n \u003c/th\u003e\n \u003c/tr\u003e\n \u003c/thead\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eFramework\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eScikit Multiflow\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eReal Time Stream Data\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1. Real time live API from Opensky.org [\u003cspan class=\"CitationRef\"\u003e30\u003c/span\u003e]\u003c/p\u003e\n \u003cp\u003e2. Real time Live API from Aviation Stack [\u003cspan class=\"CitationRef\"\u003e19\u003c/span\u003e]\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e3\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eCPU\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eIntel Core i7\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e4\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eGPU\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eA100 from Google Colab ,\u003c/p\u003e\n \u003cp\u003eT4 Tesla 16 GDDR6 memory and 2,560 CUDA cores\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e5\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eNumPy\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1.15.4\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e6\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eSciPy\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1.1.0\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e7\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003ePandas\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.23.4\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e8\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eSeaborn and Matplotlib\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.9.0 and 3.0.2, respectively\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e9\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eScikit-Learn\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.20.2\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e10\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eAnaconda Navigator\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eVersion 5.3.1; x86_64 bit\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e11\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eJupyter Notebook\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eVersion 5.7.2; 64 bit\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e\u003c/p\u003e\n\u003cdiv id=\"Sec8\" class=\"Section2\"\u003e\n \u003ch2\u003e5.1 Data Description and Metadata\u003c/h2\u003e\n \u003cp\u003eThis research used a real-time flight data collected from opensky.org [\u003cspan class=\"CitationRef\"\u003e30\u003c/span\u003e] It contains data for 583987 flights, with each flight represented by a row in the dataset. The dataset originally had 17 features, which include information such as flight number, origin and destination airport, and flight status. In order to extract more context-sensitive information, we have generated some hierarchical features from the baseline features. Original features are used to create new features that provide additional insights or context. These new features are called extended features. As a result, the dataset used here has 27 features in total, 17 original features and 10 extended features. The extended features are hierarchical features, built on top of the original features. These additional features provide more context-sensitive information and are useful for feature engineering tasks such as feature selection, feature reduction and feature extraction. This data is pre-processed and optimised with different feature engineering algorithms, only the features which are contributing to delay prediction are taken further, rest of the features are not considered. The Live Real time data is grabbed from Open sky and Aviation Stack URL, also the historical data is taken from the link provided via the same API. The historical data is used to train the classifiers and evaluate them prequentially for their possible selection in constructing the heterogeneous ensemble through classifier pool. Every epoch of live data contains about 5k instances of USA geographical area, which is used for testing the system through stacked ensemble classifiers, for prediction of flights delayed or not. In addition to this data, authors of this research have also included following data (Table 2) for experiments in order to check the feasibility of the model and observing models\u0026rsquo; performance.\u003c/p\u003e\n \u003cdiv class=\"gridtable\"\u003e\u0026nbsp;\u003cstrong\u003eTable 2 \u0026nbsp;Data Description taken for experiments\u003c/strong\u003e\u003c/div\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cimg src=\"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAkEAAACyCAYAAABFnRszAAAAAXNSR0IArs4c6QAAAARnQU1BAACxjwv8YQUAAAAJcEhZcwAAFiUAABYlAUlSJPAAADapSURBVHhe7d3/axt3nj/wpz6/p+lY+akUUzTaH0K3uPTGaek5hRTa0Tkl7HG9ldJdjkBL01FhYbs9u5WdH451vBlt2uNKsezQQDi21bjtsks4uVYK6RGJhTq+InFXcnAeEULpT5qobv+A9+eH6j0789ZIlr/GsZ4PEIne855v73nPzEszrxnHhBACRLTvxWIxtQiVSgVjY2NqMdGmVKtVHD9+XC0GTw900MX6DYKiDsBERER0f+rz9H+g9R0EERERER0k/08tICIiIhoEDIKIiIhoIDEIIiIiooHEIGgbYrEYksmkWnzPNBoNxGIxpFIpdVCHfD6PWCyGWCyGfD4PbHL8nZBKpfxlyGaz6mCKEGwzuZ2CbZfNZhGLxVCtVgNjbU8+nw/Nw3Gc0PdkMolGowHs0vz7Fdwfo/q3/B4LPOSxV8vrOM6e7Vf7QbVaDa1vKpXa9TZWyX2l0WhE9gciMAjanODOIw/6rusGatxb3377LQBgbW1NHdRhYmICpmmGyjYz/k5YXl7uWIZ+VavVPT+o3mupVApra2sQQkAIgenpaf8gL8n/f/PNN4Exty6VSuHSpUuYm5vzy27cuIFnnnkGCMwvkUiEvu/U/Pul7o9R/btSqYS+Yw+XN5PJ4Nlnnw0FYAfZX/7yFzz77LP+93K5vOlXOUQFK1Fl3cjj2LfffhvZH7ZrM8tC+xeDoD45joPr16/73xOJhH8y2i/GxsYghNhyELPd8ffSmTNn1KIDrdFooFwu49VXX/XLxsbGUCwWQ/WWl5chhEAmkwmVb0U+n0e5XO7oD41GA8eOHQPaJ5jg1dCdnP9mbHV/3MvllSfi/XT1eLfcvn0bTz/9NNDuL7quq1V6iroyXK1WcenSJbW4K/mDYbPBVz+ilo/uTwyC+lCtVnH69Gm1mO6RVCq1r67A7aXJyUk4juN/z2Qy/lWYndRoNDA5OQnLstRBKJfL/jzVX/zU2/T0NFzXDW3Dg+jatWt46KGHAAArKyt4/vnn1Spd5fN5FAqFUFmj0Yh8meO9ELV8dB8TB4yu6wKAACB0XffLi8WiX25ZljBN0//uum5oGkGVSsWvJz/BackmVKdvWZb/vVKpCNu2/e+2bYfmERzXNM3QMFXUsgghhOu6fplcb3WZ5Hc5XLaBbduR46vLHGxbVXCZerWnOh+5DJZl+XWCbQelTYLbLbisUdup13Lcj9R1r1QqoeHBdpP9Qm0T+ZGC7RbcX+S2l9MRyr6lfmzb3nD+wf6o7gOS2j/UcYLLG+wXwflIwf4tlHUVXdpLKP0eSjsHy9Vl6zZ+cNpyGsG2PkjUPhr8yH1cPX4isM3UtgMg/vM//7OjLKq/ye/q8V1uv2B/CE5LDlePw1F9TR0XgeOMOr6krm+3vk/3RufZ7D4mD9KyU0I52KiBSXCn6cZ1Xb/jBzu2ekAVEQdVNYiIGkfWkTtGr+WR0y8Wi5HTUoOc4DhyupZlRQZB3cYPHkxc1+0YR5605Dhyft0EpxU84cl1lusl21puU3kiCY4TPDkFt33U9joo1JOMejJV2yvq4C2Hye0t214GpXI8tY3lOMH+qet6KNhU5x/sp5ZlRfZbVfCkoY4jl09+l8sXNV21r0bVUZdX3R9lf46aj7psItA/g9OT/5fkPA+qSqUS2vdkQCEF20s9Bnc7xkQdm4QSmBeLRaHruj8tWS7nHdx35HTld9mH5fKo+438HrV8wXpRx/FguTwv0P5xYG6HVatVuK4LXdf9S/W6rkdeerYsq+/7xFu51WBZVijHQNd1TExMhOpI8h63vH+u63rXS61zc3MQW8xfME0Tc3NzmJub68jx6Idt25Ft8emnnwKAf7n7kUceAdo5VCpZZpomEokEEolER7KizEtaXl4OlUtRy4DA/f9uww+K5eXl0K1A13V75pjIdpycnATabS/7z5UrVwDAT3JOJpMol8toNBpd+0ivpOiNBJOr+6Hresc46vedpu6Psj/PzMyE6qHLssi2OH36NBzHwcTERNf9NZjQfpCot0jVpOj2j2//e1BUX4oqU9m2jUwmg7W1tcjtEmTbNtCerjz+yOPYRroti9pvgsdxXdcxOTmJfD6PTCbT9VxA98aBCYJ6Pd1x584dtShS8PHj2B49Ki5PaMePH0csFvO/9zpA5vP50P3xXnWlbjvvdt2+fRsAUCgUEIvF/JNtVJtHlXXTaD+uL9uj33FTqZS/DN1O5PejarXqB5GJdhKwTIp2Xbfnk3LBJM7gCUK2z+nTpxGLxVAul4HAU4KqZDKJQqHg15c/Mg7SY8fdcs0205fkCV620yCR+9/k5KR/HEX79QRqMrHsT+jzGNbL8PCwWrSneh3H19bW/EBIfZqT7r0DEwT10u8OIp8UkZ9uVyN2knxqolgshuYdFbRUq1XEYjHcvn078nHfe0H+UjZNM7T8Ub925HbY6ISSTCah6zqEEH0/VSLfA3LmzBn/l95BMz09HfqeyWQ2XNdGo+GfaNSrefIKkm3boW03NjYWeXVJbjdZz7Ztf9yo7X2QRLVHL0IIfx/d7Lj3s+XlZT84Fu1A3bIsCCH8AFy+m+nKlSuRifd7SfZp9crfZm10HF9bW/MDo36PabQ3DkwQlMlk/J1PkrfHul2O3grHcXr+6t4seRspeILrdgVKPhb+zDPP9LzytZfkwaNcLvvt4jhO5O0w+Vi167r+ryF59UHK5/NwXde/TN3tl7mUzWb9J5nQnoe8OnXQRN36unTpEnRd73p7N/jrWwYqsn/JWxay7RBoTzks2M+q1Wro9uX169f97X9QyKBSrrfsS2oA2ou8+jE2NhZ5kpfHpagfOvc79RZp8PYp2n1IBuVjY2MbXhVRrx5BuSq6FfJVJ41Gw98W3faf4GtRosjl63Ucly8TTSQSG/5ooXtATRK63wUT5YJJdMFkS5noFvyuJi+qZD016U5OS52+TLDrNr9g4qBaV01GlYL1gv+Xyanq/NVlCq5jMFEWgFhcXOwYX62jrkNUsqgctxu1bnCaRSXhW05LnV9wuWRCY3C7q+10EFTaTxiq27lXP/q3f/u30Hf5Ce4X6jaWfUTOJ9h+GyVFq/OPWlb1u0odR52mruuhbY32IUydrrpear8yTbNj2nJdosaV1Gmo30WXdpDkcmx0vLlfbZQULZQ2DLaVTBgObqtgO8nt3k87q9tGlrmuGxovqg8GxwtOZ6PlU5dHrrfVfhhFLaf9ISZ+3OhERCH5fB6Tk5PgIWLnyLd+b3RLmIj2BoMgIuqKJ+2dk81mUSgUGFQS7SMHJieIiHbe8vIyXn311cjcDOqf4zhoNBoMgIj2GV4JIiIiooHEK0FEREQ0kBgEERER0UBiEEREREQDiUEQERERDSQGQURERDSQGAQRERHRQGIQRERERAOJQRARERENJAZBRERENJAYBBEREdFAYhBEREREA4lBEBEREQ0kBkFEREQ0kBgEERER0UBiEEREREQDKSaEEGphlFgsphYRERERoc9QYt/pOwgiIiIiOkh4O4yIiIgGEoMgIiIiGkgMgoiIiGggMQgiIiKigcQgaJscx0EsFkO1WlUHdfA8D47jYHR0FPl8PjQsm81iaGgInueFyg+yer2OoaGhjrYgIiLaC7seBDUaDeTzeQwNDamD9rVqtYqpqSm1uMMDDzygFnW1vr6OO3fuYHV1VR0ETdPUogPv0KFDahHtsX72z6WlJSSTScRiMT9ojQrWPc/D1NQUhoaGEIvFkM1mI+sRbUa//a9eryOVSiEWi7H/Uf/ELqrVasK2baFpmtjlWe040zSFpmmi2Wyqg7alUqkIAMK2bXXQgdZsNgdunfe7fvbPUqkkNE0TuVxO2LYtdF0XAIRpmmpVYZqmMAxDuK4rms2mMAxDGIahViPqW7/9r1arCU3TxPz8vBBCCNd1/f6308dwOliij3w7zDTNrgfZ/ch1XQFgV4KVQQ2CbNseuHW+X/TaP03TDJ1Ems2mfyKq1Wp+ebFYFABEpVLxy+R+xO1OW9Vv/zMMQ6TTaf+7GOBjLW3Ort8Oux9dvHgRtm0DAC5cuKAOpk2q1+uYnJxUi2mfazQaOHPmDOLxuF8Wj8cxMzMDAPjhhx/88nfeeQeapmFsbMwvSyQS0HUdly5d8suI+tVv/2s0GlhdXcUTTzzh1wOAsbEx6LrOYzj1xCBI0Wg0cO3aNUxMTMCyLLRaLTiOo1YD2nWnpqaQTCb9xGi1bGpqCrFYrGd+UbVa9ROju5XJvI2hoaHI5VlYWPDvmyeTSSwsLKhVIi0tLWF0dNS/357NZtFoNPzhm10O1dLSEh5//HEAwOTkJGKxmH9Pf2FhoSNJfGlpCZlMBqlUys8xkcsm5+c4jr+u2WzWH1eqVquhdZqammJuwBYkEglkMhm1GA8//DAQyOnyPA+rq6s4duyYUhNIJpNwXTfUp4j60W//+/bbb4F2zqUqmUyi1Wpx/6euGAQpLl68iFdffRUA8NJLLwEApqenlVo/+uGHH9BqteC6bmTZRx99hMceewymaYbGi3Lz5k20Wq3Isnw+j+HhYVy9ehW6ruP06dOhk8rU1BTq9Tq+/PJLNJtNPP/883jttdc2DFIWFhbwy1/+Ev/6r/8KIQSuXr0Kx3FgGEZo+v0uR5Tx8XFUKhUAgG3bEEJgYmLCP2AFk8Q9z8MDDzyAxcVFeJ6Hd999F3/7t3+LUqkEtJ+gy+fzuHPnDq5cuQLLslAoFEIB39LSEn79619jcXERQgj87ne/w+zsLM6dO+fXoe357LPPYJomRkZGAAC3bt1Sq3SQJyqi7VL739GjRwEAi4uLSs2/+uabb9Qioh+p98d2Q6+cg/2k2Wx2JEPLZS+VSqG6km3bHbkQsqxYLIbqih73qaPaSJYFl0fNvZB5F+p9cwBC13W/TCXHk4mEkpy+ZVl+WT/L0Uu3de5WLhNsg2SbBreDXM/g+Lqud2wrwzAEAOG6bqicfhTV97qRORnBfAy5HdVkVRGYdj/9hGgjUf1PCCFyuZx/3Go2m6LZbIpisdgz6Z9IMCco7PLly7AsK3QP+syZMwCA9957L1CzP/Ky7XYFl0ed5ueffw4AOHLkiP9o6JEjRwAgdIVKJcd79NFHQ+WZTAaapqFQKITKscFy7LTgvIKCrySQda5fvw60c49c18XJkyf9tojFYv7VJl6N2L53330XMzMz/q9wor3Urf+dP38euVwOjuPgyJEjeP3117G+vo5WqwXDMEJ1iYIYBLV5nocLFy5gdnY2dAI9ffo0AKBcLvf1QsS9Jm8rtZ/06/h0E3X/XIrK7dhIPp8PtZvM/dlLMlGyUql0tIMQIpS0S5vnOA6Gh4c78jQeeuih0Pco8pYF0VZ163/S+fPncffuXQgh4DgO6vU6AOA3v/mNWpXIxyCo7fLlyzh27FjHiVMIgVwuBwB4//331dHuucOHDwPthGzVRjlBAPD111+rRcAWXt749NNPw7bt0Ofpp59Wq+0qmSgZdf+/0WjsyyD2flGv13Hjxg2cPXtWHeQ/BbaysqIOwsrKCnRd73plj6gfvfpflGq1ikKhANM0uwZNRGAQ9CN5FWhubk4dBAB44403oGkaFhcX/V8X+8VTTz0FAEin06EnIDzPw3//938Haoa9+OKLAIDf//736iCsrKzgrbfeUot7Ghsbw8TEROiz11deRkZGoGkastlsx3a6ePHini/PQVGv1/Hxxx933T8AYGZmBq1WKxRo1ut1tFot/5Fmoq3op/8F1et1nDp1CoZh4MMPP1QHE4XsehDkeR7W1taALlcr7jXP83Du3DkMDQ35V1VU8Xgcuq4DAF555ZXQety+fRtQrqjIss8++8wvk2S9r776yi/zPM8PYIIn76gyOf5f/vIXoH3itywLq6urePLJJzE1NYV8Po9UKoWXX37ZH0+VSCSQy+Xguq7/ennP85DNZqHremjcfpajF3kr5JNPPsHS0pJ/m0xesVHbYmVlBWtra6GgLqqd5fIE229ubg6tVgsnTpzwnyZLpVJ44YUX/PHorzbaPx3HwT/8wz/g8OHDyOfzoU8qlfLrZTIZGIaBX//61/A8D41GA6+88goMw+Avcdqyfvsf2seDqakpPP744zh27BiWl5d5BZI2pmZK7yT5RE/woz4JdK/Jp1fkJ+opFnUdZD11XNu2O8qCT8xEtUfU9OXTNpudtnyTqmEYHU9PdDM/P++Pp2ma/3SFtNnl6EY+vWG23wCrtoVpmpHrHdXOpml2jI9AVy6VSv4TYbquRz6lR9H9Mbh/lkqljuHd6or2kzuWZQm0+1Iulwv1JaLN2Ez/k2XpdJr7O21KTPTKniUiIiI6oHb9dhgRERHRfsQgiIiIiAYSgyAiIiIaSAyCiIiIaCAxCCIiIqKBxCCIiIiIBhKDICIiIhpIDIKIiIhoIDEIugeq1SpGR0cRi8UwNDS0539t/SCpVqtIpVJsQyIi2jQGQXusXq/jzJkzWF5eRrPZRCaTwYULFyL/bhP1Vq1W8dFHH6FcLquDiIiINsQ/m7HHMpkMvvvuOywvL6uDaItisRhs28bExIQ6qKelpSU88MAD/OvyREQDileC9ti1a9fUIrpH3nvvPbWIiIgGCIOgPdZqtdQiugcWFhZ4G42IaMANfBBUr9cxNTWFZDIJAMjn8xgaGsLQ0BCmpqbU6h1JzVNTU/A8DwDQaDT8aVWrVUxNTSEWi2FqagqpVAqxWAwAUC6XEYvF/O8A4HkestkshoaGEIvFMDo6CsdxQsMXFhYwOjqKfD4Px3EwNDSE0dFR1Ot1f1y0T/ByOnId6vW6v9ypVKojB2lpaQnJZNJfrkwmg3q9Hjnver3ur08qlfLXX2o0GqF1SaVS/rSCdTKZTGh+6jJFqdfrofGitpFsS1knmUyGEqenpqbw2muvAQCOHz+OWCyGarUKbNAORER0wIgBV6lURDqdFgCEbdvCsixh27bQNE0AELlczq9bKpWEYRjCdV0hhBDz8/MCgLAsSwghRK1WE5Zl+WXFYlGYphmaBgBhmqb/XQghms2m0HVdWJYlms2maDab/nTkuK7r+vNLp9Nifn5eWJYlDMMQlUpFGIbhr0OxWOwosyxLVCqV0DSkSqUiAIj5+fnQd1lHnXculwtNS66/rKtpmj8t13WFrutC0zTRbDZDZZVKxZ+fpmlC13V/OlFqtZrQNM1vJ9d1Q+somaYpdF3321Ju31qt5texbVsA8JdB9NEORER0sAx8ECQCJ8RSqeSXyZM5AP/kret6qI4Qwj8Jy8BITqtYLIbqSYgIgizLEpqmhcpEe37BacuTcjDokOR85bKKQP1gECbayxxchlKp1BEkmKYZqtNtWrquh+rJwCwol8sJTdP89bAsq2M6uVyuZ7uJ9jKp067VagJKEGQYRmSQFwx4ooKgftqBiIgOjoG/HRY0Pj7u/z+RSCCTyQAAbt26hXq9Dtd1cfLkSf9WSSwWw+rqKgDg22+/9ccFgIcffjj0vRfHcXDs2DG1GP/8z/8MAPj0009D5Y888kjoe1A8HleLcPjw4dD3eDweyocZHx+HEAIjIyNwHAepVKprvow6LXkbUSoUCvjHf/zHUNn58+dx9+5dJBIJoL2+s7OzoXacnZ0FANy5cyc0rtRoNFAulzumPTIyEvoOADdv3oTjOP7tyVOnTqlVIm2mHYiI6P7HIKiHYLDxww8/AAAqlQraV9BCn+08Zt0tWfrRRx9Vi3aNzIW5ceMGbNuGaZpqlR3TarVg23ZHGwohuj7mrgaZvcicoOeffx6PPfYY/vCHP6hVutrLdiAionuLQVAfDh06hEOHDgEAvvnmG3UwGo2Gn1i7VSsrK2qRT736stMcx8HJkyfxxz/+EXNzc5FXVzbjk08+UYvgeR6WlpYAAJqm4fbt22oVoL0svayvr6tFIZ7nIZVK4e7du1hbW0Mmk8EDDzygVou00+1ARET7G4OgAPUpoOvXr0PXdYyMjGBkZASapiGbzXbUu3jx4rauBFmWhVar1REAfP3119A0DS+++GKofKdduXIFmqbtyEk/nU5jdXW1Y13OnTuHJ598EgDw/PPPo1AodNRZWFjA0aNHQ2WSLF9cXOx4Gi3o1q1bWF1dxYkTJ9RBG9rJdiAiov2PQVDAwsKCf4KV75G5cuWKP3xubg6tVgsnTpxANptFPp9HKpXCCy+84NeRVzg+++wzv0ySV4vW1tZCJ/I333zTD7BknWq1irfffhtzc3N+no+8CnX9+vWOQOD69euAEsjJ+l999ZVf5nke1tbWgPYVLAB48MEH0Wq1kM/n/Uf719bWsLKygqWlJTQaDXz//fddp7WysuIvz9tvvw1N03D69OlQGz3yyCP+egTrZDIZ5PN5ZLNZ3Llzp2sAEo/HYds2XNfFL37xC3/ZZSB16dIlLC0t+VfsPvjgAywtLcFxHLz//vtAe5vIq1E//elPAQAfffQRHMeB4zh9tQMRER0gaqb0IJJPCgUfjZePnqvkY/IAhK7roaeZTNMUAPxP8KkidZg63HVd/1FuOf/gk2hyGYMfSS23bTuyvnxKSi2r1Wr+k2imaQrXdf0nuorF4qamJdpPbMn11XXdf+Q8qFar+euraVrH02LdzM/Ph5ZVPjafy+X8p8/kk2byUX35pJ9pmv7Tc81m019GOe+N2oGIiA4W/u2w9gsSJycnwaYgIiIaHLwdRkRERAOJQVAgz0XmixAREdHBN/C3w2KBv98FAKZpYnl5OVRGREREB8/AB0FEREQ0mHg7jIiIiAZS31eC1NtGRERERADu26er+w6CiIiIiA4S3g4jIiKigcQgiIiIiAYSgyAiIiIaSAyCiIiIaCAxCDpAGo0GFhYWkEwm/b9Gv9ccx7mn8yciIurXrgZBS0tLSCaTiMViGBoaQj6fh+d5arV9xXEcpFIpxGIxxGIxJJNJZLNZNBoNteq+88MPP6Ber8N1XXXQnnAcB1euXLln86fN2e7+6TiOv5/ITz6fD9WpVqsYHR31h4+OjjJApk1rNBrI5/MYGhpSB/mC/TmZTGJhYUGtQtQp/Efld06pVBKapolcLids2xa6rgsAwjRNteq+UKvVhGEYQtd1USqVQuXpdFoAEMViMTTOflQqlQQAUalU1EF7olKpbHn+uVxOLaJdshP7p2EYwjTN0KfZbPrDa7WaACDm5+f9svn5eQFA1Go1v4yol1qtJmzbFpqmiW6nrFKpJNLptKhUKqJUKgnDMASA0LGcKEp0j9oB6gGx2Wz6B9r9dgBsNptC0zShaVpomYPul0BoO0HITtjq/OU2oL2x3f2zWCyGgpsouVxOGIahFgtd14Vt22oxUU+maXYNgtS+6LquACDS6XSonEi1K7fDGo0Gzpw5g3g87pfF43HMzMwA7ds2+8m5c+fQarXwu9/9LrTMQbOzswCAbDa7qVsG1J/XX38drVZLLaZdsN390/M8TE9P4+2330Ymk+l5e8t13Y795e7duxgeHg6VEW3H2bNnQ98TiQQAQNf1UDmRaleCoEQigUwmoxbj4YcfBgAcOnRIHXTPeJ6HQqEAAHjuuefUwb5EIgHTNNFqtfD555/D8zw4joNMJoN8Po96ve7nEo2OjqJer4fG9zwPU1NTGBoa6siN8DwPCwsLGB0d7ZhWKpUKnUQ8z0M2m/XzOFKpVGRbS9VqNZSzEcx3kp/gSSxY3o+lpSU/52NoaAgfffSRWiW0PnLdl5aW/OGjo6NYXFwEAvOXZB6AnD6D0O3b7v755ZdfQtM0tFotLC4u4vjx45Hb5ec//zlarRZSqZS/P0xNTSGTyUTOn2inVKtVaJqGN954Qx1EFLIrQVA3n332GUzTxMjIiDronrl165b/f/nroRs5/MaNG1hfX8cDDzyAxcVFrK+v45VXXsGzzz6LdDqN1dVVnDhxIpRMnUqlMDw8jLt37/qJw8ePH0ej0cD6+joAYHV1FV999RU+/vhjTE9PY35+HuVyGefOnfOnc+7cOdy9exfNZhP/93//h0Qige+++84frhobG4NpmrAsC67rYnl5Ga7rwjAMAECpVMLY2Jhfv1KpQNd1NJvNwFSiOY6DkydP4pVXXoEQAlevXoXjOKE6nufhxIkTePDBByGEgOu6aLVa+OUvf+nXuXnzJkzTBH681u3/DZp8Po/JyUl88cUXEELgrbfeQqFQwKeffuqPSzun3/1zfHwcN2/ehBACxWIRuq6jUCiE+ikAjIyMoFgs+vtDJpPB8PAw5ubmQvWIdornecjn8zh16hSuXr3a9co+kU+9P7ZbZM5BP/kGe0nmsPTTFLZtdySPyu/B/IpcLicA+Im+xWKxIzdCJjBbliVEYDnU5GBd10PzM02zo46chlBycprNprAsKzI5UNZTczNs2+64vx5F5vCoyyITX2VOkOu6QtO00DRlOwZF3e/P5XIdeUJRy0zbt539s9ls+omoUePL7S37e7e8O6Jeoo4RKtnPAAhN0/Z9Difde3t2Jejdd9/FzMzMhr8y9zN5xeZv/uZvQuXPPvts6BfHyy+/DAD4r//6LwDAn//8Z6yuroZuNZ08eRJo52cEHT58OPQ9mUyGvv/qV7/C7Ows8vm8P27UL+uvv/4aTz75JN58802Mj4+rgzE2NoZ0Oo0LFy6EluGTTz7Biy++GKob5fPPP0er1cLf/d3fhcofffTR0PdEIoG7d+/i7NmzqFaryGazmJycDNXp5vz587h7965/u1BtC9o529k/4/E4PvjgAwBAuVwODXMcB7dv30az2UQ6nUahUOi4xUu0U+TVZtu20Wq1Im/TEgXtSRDkOA6Gh4f3ZR5A8FaQmsejkkHNY489pg4KUW+rfffddzBN07/VE/wsLy+H6m5kfHwclUoF169fh67rXd9hJN8XlM1m1UG+2dlZtFotXLx4EWhvp9HRUT+gU/OJYu2cIgC4c+dOaFq9yJygmZkZvPDCC7BtW63SVT6fx5NPPon19XVcu3ZNHUw7YCf2z5GREf8Wq7S0tITp6WnMzc0hHo/DcRxYloXV1VVcvnw5VJdopyQSCUxMTPiBUDDlgUi160FQvV7HjRs3OrL39xPLsoCIX7FBnuehXC5D07SeCdRBDz74oP/v2tqaOhhon4A2a2xsDMvLy6hUKrh58yYMw+j4tfPSSy+hWCyiXC53DYQSiQQsy0KhUEC9Xsf09DTefPNNf/hDDz0E27ZDnzNnzoSm8f3334e+q+r1Ok6cOIG///u/x/LycuRVqW6y2SwuXbqEL7/8EhMTEx3BJW3fTu+fMrcLAN57772Oq3dzc3MwTRPXr18PlRPttH6uaBPtahBUr9fx8ccfR96u2U9++9vfQtO0jltDQTLp8w9/+ENHst1XX30V+i6ftvrZz34GADhx4kTkVZl6ve7fYuvX1NSU//+xsTEsLi52/bWTyWQwPz+PQqHQMW9JBj0nTpyAYRihQEP+ogp+5NWCn/70pwCAf//3f/frRymXy2i1WnjqqafUQRsqFAowDKOjvWln9Lt/dtsngur1OuLxeOh2WrfgX72dTLQb/vd//xeapuHo0aPqIKK/UpOEdkqxWPRfiqZ+NvNW2r1Sq9WErusdyXTyjdFquSQT8IKJwPJNulIwcdQwDGHbtsjlcqE6MlE6+HIvmawafImjaZrCsizhuq4QgaRT+d2yLAHlLb0yobBbUqpM5N7sCw7ldHO5nGg2m6LZbPrTkm9vlYnS8rvc/mi/eFLOU443Pz8vcrmcn1At271UKvl1TNOM3BbUv373T7n9gv1J0zRhmqbf5yqVSsfDAbJc7XfyTdVRyfpE3QRf5in7nSSPFcHjouyT7Ge0kV0JguQJvdtnvz7d02w2hW3bfsACQOi6Htq5VPIEL0/s8okp9YQgn9TS2q9+T6fTfp3g0zPyE3xqLViWTqdDy2cYhh9IyGWQH9M0u04nSAaAmyWDHrlO8kk0XdfF/Py8HxjJ5dJ1XVTar7WH8iRcVBBaLBb9acs2lX/aJOopJOrPZvZPWTd4MpHBKAJBvdrfpUql0rW/EvUj6vionkPkG/3RxzGbKCgm5EtZaEtisRhs28bExIQ66L4xNTWFxx57bFuJsURERPebXc0Jov3P8zwsLi4yACIiooHDIGgb5CP1n3zyiTpoX3McB7FYDMlkEj/5yU/8vxlFREQ0SBgEbVE+n8fjjz8OtP/cRUz5G1z72dGjR6FpGtB+ZJlXgYiIaBAxJ4iIiIgGEq8EERER0UBiEEREREQDiUEQERERDSQGQURERDSQ+k6MjsViahERERER+gwl9p2+gyAiIiKig4S3w4iIiGggMQgiIiKigcQgiIiIiAYSgyAiIiIaSAyCiIiIaCDtahC0tLSEZDLp/8XyhYUFtco9l0qlEIvFun5SqRQWFhbgeZ46qq9arSKTyWBoaAixWAxDQ0PIZDKRf1A1n893zKPbJ2r8arXaUS8WiyGfz4fqRa2X5Hkestmsv7yjo6NwHAee53VMJ0rUtIOfVCqFbDYbufzdpFKpyDbebnt5nodUKqUWU5vjOB3t2E8fCHIcJ9T/VY1GA/l8HkNDQ+ogor7004fq9Xro2JTNZiOPKUQhYpeUSiWRTqdFpVIRpVJJGIYhAIhSqaRWvedc1xWapgkAolKp+OW1Wk2k02kBQBiGIZrNZmi8ZrMp0um00DRNzM/P+8Nd1xW2bQsAwrKs0Di2bYv5+flQGQBhmqb/vdlsCsuyhG3boXpBuVxOABC6rncslyTXKzi/ZrMpDMMQhmEI13X9enJ5e80zqFubyWnJYblcLjRelEql0nXe220vuV7BZaS/MgxDmKYZ+nTrT6pmsylM0/T7eVQb12q1UH8g2qx++lCtVgsd61zX9Y9z/fZnGkzRPWoHqCcu13UFAJFOp0Pl+4U8mEcdyOUwdZ16jSPabaAGQsViMVRHRJzUpaiTuiQDh6jxgtThcplkABRk23bPeap6rb88KHULboIsyxIAhKZpHQes7baXrutd6w+6YrHY0af7JYNpTdNErVZTB3eQfYVoq3r1IcMwOs4tvX5cEUm7djvs7Nmzoe+JRAIAoOt6qPx+8OyzzwIA1tfX/TLHcVAul5FOpzE2Nhao/Vdnz56FrusoFAr+rZpMJqNW62piYkIt2ja5Dt9++606CC+//LJatGUjIyO4evUqAGBychKNRkOtArQvc1+7dg2WZaHVauHTTz8NDd9OezmOg+effx66rqNcLqNer4eGDzLP8zA9PY233367663bXn7xi19gdXUVX3zxBUZGRtTBRHum0WhgdXUVTzzxRKh8bGwMuq7jwoULoXKioF0LglTVahWapuGNN95QB+17t2/fBgAcPnzYL3vnnXcAAD/72c/8siivvvoqAOCjjz5SB90TL774IgDg1KlTWFpaCg2Lx+MdgcR2jI2NwTAMAMDnn3+uDgYAXLx4ETMzM3jzzTcBAL///e/VKls2PT2NN998EzMzMwCwL3PS7pUvv/wSmqah1WphcXERx48f7zuHQv4AsCyLARDdc/IHXfBHqpRMJtFqtfrq1zSYdj0Iksm2p06dwtWrVxGPx9Uq+5ZMxisUCjAMww8gAGB1dRUA8PDDDwfG6DQ8PAwAuHbtmjronkgkEigWiwCAkydPIpVKbfoqwGaMjo4C7aRFVaPRgOM4eO6555BIJGCaJlzXheM4atVNcxwHyWQSiUQCzz33HDRNQ6FQ6HpFatCMj4/j5s2bEEKgWCz6VyzPnTunVu1w5coVAMAjjzziJ6IODQ1hampKrUq0644ePQoAWFxcVAf5vvnmG7WICNiLIOjIkSOYnJxEq9XCqVOnduQEt5uOHz/uP12g6zomJydhGAYWFxe3FMDJIMl1XXXQPZPJZLC6ugrLslAul3H8+HGkUqnIQGW7HnnkEaAd8KguX76Mt956y2/X6elpIHCVbTveeecdf3rxeBxvvfUW0J4nhWUyGXz55ZcwDAOFQmHDflAul6FpGoaHh/Hhhx+i2Wwik8lgdnaWgRDtuXg8jlwuB9d1/auZnufBcRysrKwA7dvzRFF2PQgSQsB1Xdi2jVar1fcl93ulUqmgnTCOWq0G27bhui4Mw9hSAPf9998DgH9baL9IJBKYm5uD67owTRPlchmPP/74ltaxF3krUeaESZ7noVAohPKQ5D381dXVbV2dkuMGc7XkVbxCobCv+9+9Eo/H8cEHHwDtIKcbGSAdO3YMmUwG8Xgc8Xgcc3Nz0HUds7OzbF/ac+fPn0cul4PjODhy5Ahef/11rK+vo9Vq7btjL+0vux4EoX0CnJiY8AOhW7duqVX2pZGREUxMTGBubq4jgDNNEwDw9ddfK2OF/c///A8QuC20Ger7W6LewbJdiUQCy8vLKJVK0DQttI5R7wPabHBy8+ZNAMAzzzwTKr98+TJarRaOHDkSmr68YibzeLZiZmYGq6uroenKhPxWq8WrQV2MjIxseMLodfs3nU4DwH2zf9PBcv78edy9exdCCDiO4wfsv/nNb9SqRL49CYKkYE7N/UQ+oRQM4M6cOQMA+NOf/hSqq/rkk08AAC+99JI6aEO2bXd8pIceeghoX1HpRf1Vn8/nI29NjY+P+09oBddRnb+cbz+q1SpWV1ehaRqee+45v9zzPFy4cCF01U1+XNeFpmlbfpqrWq1iZWUFzWazY9qlUgkA+LTIBmSAHyUej0PTNKytramD/AcHDh06pA4i2lPVahWFQgGmaW7qCVMaQOoz87upVCpFvgtmP+j1zhv5vgn1/TpynG4vgCwWiwJ9vDBwq++xke/AiXrnj2i/E0h9d4Zt212XR76AsZ/3vogN2iz4niD1XT+2bXe8RDJIvuBQXXapV3uZptkxv6Bu73yiH7dZVLuq/avbCyjT6bQwDCNUJvV6xwtRP/rtQ/LYwxclUj827lFbIN8mbFmWfwCtVCrCNM2uAcO91O3tx81mUxSLxa4v/Qu+Mde27cg3RncLOCQZKPUKZroplUoC7bdZB9u1VquJXC4X+SI7uVyWZYWGyfXsFZwEdWsz9Y3RakBSqVRCb3aNItcrKljp1V7yRZDqOgfJQE/TtI6T+CDRNE2Yptmxf6onDdmmG711fH5+PrK/yfobBexEvfTTh+RxD+0fSWpfJoqyK0GQaP8qlCcrXddDAdF+IoOYbh9N00Q6ne4ZvBWLxdB05DgbnWTVecmddzPkn/aQQUewvaMOAjJYk3/WRI5jGEZHwNHNRm1mmqbI5XId2zt4Ra3busogLaqeWh4cpo6nBqyiy3JH1RsE8mSB9rYPBvFBMiBV+3+z/adK5DTS6XRkAKRul0Fuc9qafvpQsB+qP7yIeomJHzsQERER0UDZ08RoIiIiov2CQRARERENJAZBRERENJAYBBEREdFAYhBEREREA4lBEBEREQ0kBkFEREQ0kPp+T9Bu/PFOIiIiuv/1GUrsO30HQUREREQHCW+HERER0UBiEEREREQDiUEQERERDSQGQURERDSQGATtgWQyiVQqpRZv2U5P737geR4cx0EqlUI+n1cHExERbdqeBUGe5yGZTN43J7BqtYqpqSm1eEs0TVOLtmWnp3c/WF9fx/r6OsrlsjqItqHRaCCfz2NoaEgd5PM8D1NTUxgaGkIsFkM2m4XneWq1DTmOg0wm408nqFqtYnR0FLFYDLFYDKOjo6hWq6E6NJiWlpaQTCYRi8UwNDSEfD7ftf/105+JQsQeSafTAoCwbVsdtC+Zpik0TRPNZlMdtKdyuZxatO+VSiVRqVTU4m1rNpv3VR/a72q1mrBtW2iaJnodCkzTFIZhCNd1RbPZFIZhCMMw1GpdNZtNYZqmACAsy+roG7VaTQAQ8/Pzftn8/LwAIGq1WqguDZZSqSQ0TRO5XE7Yti10XRcAhGmaatW++zNR0J70lPn5eWHb9n1zAnNdVwC458tbqVQid/b9zjTNjhPdTrnX2+QgkgFKlGKxKACEtqfcP/rZDjJo0jSta0CTy+Uigypd1/uaBx1cpmmGfog2m00/EOrWn3r1ZyLVrt8Oq9fr+NOf/oSJiQl10L518eJF2LYNALhw4YI6eE94noczZ86oxfvewsICb1kdIO+88w40TcPY2JhflkgkoOs6Ll26FKob5Re/+AVWV1fxxRdfYGRkRB3sc1234xbH3bt3MTw8HCqjwdFoNHDmzBnE43G/LB6PY2ZmBgDwww8/BGoTbc2uBkGe5+GVV17Bhx9+qA7atxqNBq5du4aJiQlYloVWqwXHcUJ1UqmUn7sQi8VCeU6yTCYuyzwINZE5eJ87Foshk8mgXq8D7WX4yU9+Atd1US6X+5oeAOTzeX+aMv8qeGIJjut5HrLZrF9XznsjwfGGhoaQSqWQyWQAAFNTU3jttdcAAMePH0csFvPzOnqtb3Da6josLS2F6qiC2yHGP+2yozzPw+rqKo4dO6YOQjKZhOu6aDQa6iCf4zgol8uwLKtnAPTzn/8crVYLqVTK7xNTU1PIZDJ+36LBk0gkIrf/ww8/DAA4dOiQOoho89RLQzspnU6HLln2ewn9XrIsy1/GSqUiAAhd10N1ms2mn+NUKpVCw1zXFYZhiGazKZrNpj+N4G0tWSZzIOT3dDodmNKP7RUcr9v0RLutZd6GCORUyNsMwXENwxC5XE6USiX/nru6jt1YliXS6bS/fpZlhZZF3vYM3j7pd30NwxCWZfnTlm0cnJbah2zb3tXbb4Og2+2Dbn1NBMbp1e6yjtxGAPz8DpW87aZpmkin06H8IKKgXC4X2Selbv2ZKMquXQlaWFjAiRMnev4C3G/kY9gvv/wyAGBsbAymacJ13dAViXg8jtnZWQDAf/zHf/jlAPDpp5/iX/7lXxCPxxGPx0O3EaTvv/8eAPDUU08Bgfl89913Ss2wbtNzHAeLi4v44IMPkEgkAABnz56FZVlYXV2F4zgd454/fx7j4+MYHx9HJpOB67qBKXbXaDSg67q/fnNzc/48u+lnfR3HwerqKn7729/60/6nf/qn0PgqeQVueXk5sl3o3iqXy9A0DcPDw/jwww/RbDaRyWQwOzvb8eRlJpOBbdtotVpYXFxEvV7vuD1G5HkeFhcX/XQFou3alSCoXq+jXq/j7Nmz6qB97fLly7AsK3QPWublvPfee4GaP16qtSwLhUIhdEvg+vXrGB8fD9VVjY+PQwiBkZER/90328mj+fOf/wwAHQGnbP8rV66EyoPrBwCPPPJI6Hsvv/rVrzA7O4t8Pu+v99zcnFotpJ/1vXLlCkzTDC2bHE9tz/X1daRSKQwPD99XuWaDRN7WOnbsGDKZTCho1nUds7OzHbdqb9++jWaziXQ6jUKh4N+2JZLeffddzMzMdBzriLZqV4KgyclJFAqFyHyNycnJUK7IfuF5Hi5cuIDZ2dnQMp8+fRpo/6pVl1kGGRcvXgTaB/J+k5lljsyNGzdg2zZM01Sr9K3bFaStHijUnKfg9hofH0elUsH169eh6zqy2WzPvBBpJ9fXdV2srKxs+X011J+HHnpILepw9OhRtQgI5G1ESafTAIBbt24B7b4xPT2Nubk5xONxOI7jX8W8fPmyMjYNKsdxMDw8HJknRLRVuxIEnTlzBrZtd3wAwDRN2Lbd1wF2L12+fBnHjh1D+7UBoU8ulwMAvP/++6FxRkZGYJomCoUCPM/DlStX+tpBHcfByZMn8cc//hFzc3NbDlZUaqKx9OCDD6pFPUVtv+D2Ghsbw/LyMiqVCm7evAnDMHoGI/2ub7lcjpyOoySmP/HEE/jiiy+AdsAWNQ5tn3wKbGVlRR2ElZUV/7ZolHg8Dk3TsLa2pg7C4cOHgUBi63vvvYdkMhmqMzc3B9M0cf369VA5DaZ6vY4bN27cd3cX6D6gJgntJjWpdb9oNptC0zQ/qVglhyPi3RTBZONu66Yml5rtFzEGmabZkeynjtetXCaVWpYVqidfQqcmFqvTlMnM/VCTWuU7Y+Q8ohKj+1lfmcitrkOxWAwlnwf7kFw/mYhOW9MrkTTqPUGy3YvFYqiuuv9E9QURSOIPfo9KzN8oAZYGQ61W6zju9NKrPxOp9rSnBE9g+4V8wknX9Z4nUsMw/BOuerCXJ/mo8WWQEJx+8O3ZlUpF5HI5oeu60DRNlEolf/ryJXOlUskPDqKmJ+si8ASWfEotGFQ0229cVseNegqrG9M0hWVZ/jLKE538XiqVBNrBTLFYFMVisa/1bbZfqod2kGbbdseTZ/KEHHyqTM4/arvQxpqBl891az+j/YboZrPp96tgECMCQWzwqS65TYPbZn5+vuPFifKHhHwyUATeFKw+fUmDpVgs+i/NVD9RAXI//ZkoaOCDIPmrQX6iAoHg8Kh6pVIp8peKPLgHP6L9yyb4+nfXdUUulxOapoV+XReLRf/x9Vqt1nV6or3zy2mgHegE2zpq3KiyjbaP/BUv6xuGEWqLZuBPJMg26Xd9ZUCKwKPU8qQog53gR0Rsm42Wn/4qqk2j2q/XdpFk8KsGLcFx0Q5g1aupot0/e/UrGjyyT3X7qH213/5MFBQTP55IiIiIiAbKriRGExEREe13DIKIiIhoIDEIIiIiooHEIIiIiIgGEoMgIiIiGkgMgoiIiGggMQgiIiKigdT3e4LkH0AlIiIiCuozlNh3+g6CiIiIiA4S3g4jIiKigcQgiIiIiAYSgyAiIiIaSAyCiIiIaCAxCCIiIqKBxCCIiIiIBtL/B5jjdAYzZsB2AAAAAElFTkSuQmCC\"\u003e\u003c/strong\u003e\u003cbr\u003e\u003c/p\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec9\" class=\"Section2\"\u003e\n \u003ch2\u003e5.2 Experiments\u003c/h2\u003e\n \u003cp\u003eWe have conducted several experiments with real time data streams for binary classification, to predict flight delays. The scope of this research is limited to only binary classification. Further it can be extended for multiclass classification as well. The major goal of this research is to build the stacked ensemble such that it will find a golden midpoint between the expected accuracy of binary classifiers over a real time continuous data stream and the computing power requirements as well as time \u0026amp; memory requirements simultaneously. This is a very first requirement of any real time stream data analysis system, since the stream data that too if it is live streaming has to undergo data preprocessing, data pipelining and various feature engineering aspects almost on the fly, at least in a single run. Having said this, the system of this type of data analysis should have the prime capacities of giving highest performance consuming as less memory as possible and with very less computing power as far as possible, so that it will not be resource and memory savvy. (Table 3)\u003c/p\u003e\n \u003cdiv class=\"gridtable\"\u003e\u0026nbsp;\u003ctable id=\"Tab3\" border=\"1\"\u003e\n \u003ccaption language=\"En\"\u003e\n \u003cdiv class=\"CaptionNumber\"\u003eTable 3\u003c/div\u003e\n \u003cdiv class=\"CaptionContent\"\u003e\n \u003cp\u003eDesign of Stacked Ensemble \u0026ndash; Selection Criteria\u003c/p\u003e\n \u003c/div\u003e\n \u003c/caption\u003e\n \u003cthead\u003e\n \u003ctr\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eSr.\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\u0026nbsp;\u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eComponent Classifiers\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eParameters/Criteria\u003c/p\u003e\n \u003c/th\u003e\n \u003c/tr\u003e\n \u003c/thead\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" rowspan=\"4\"\u003e\n \u003cp\u003e\u003cstrong\u003e1\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" rowspan=\"4\"\u003e\n \u003cp\u003e\u003cstrong\u003eTier 1\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eHoeffding Tree (HT)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" rowspan=\"4\"\u003e\n \u003cp\u003ePrecision, Train_time\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eExtremely Fast Hoeffding Tree (EFDT)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eNa\u0026iuml;ve Bayes (NB)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eHoeffding Adaptive Tree (HAT)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" rowspan=\"3\"\u003e\n \u003cp\u003e\u003cstrong\u003e2\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" rowspan=\"3\"\u003e\n \u003cp\u003e\u003cstrong\u003eTier 2\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eVery Fast Decision Tree (VFDT)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" rowspan=\"3\"\u003e\n \u003cp\u003ePrecision, Train_time\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eVery Fast Decision Rules Classifier VFDRC\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eNa\u0026iuml;ve Bayes (NB)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003e3\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003eFinal Tier (Metaclassifier)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eGB,LR,ARF\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eAccuracy,AUC-ROC and Quick Response\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003c/table\u003e\n \u003c/div\u003e\n \u003cp\u003eWe have evaluated each of the candidate classifiers for their selection into classifier pool of stacked ensemble layer (viz.1,2,3 etc.) and keeping a threshold value for mean precision and train and test time required for each classifier, our model finally selects the classifiers dynamically. Figure 3 (a,b) shows the example execution results of a random epoch execution over a part of the sliding window of real time flight data stream. The similar process is adopted in order to select a metaclassifier, there are directions received from subsequent literature surveys to go for GB-Gradient Boosting, LR- Logistic regression and RF- Random Forest and some variations thereof. We have selected ARF- Adaptive Random Forest since its performance is undoubtedly good when acting as metaclassifier over a big data stream of dynamic nature [\u003cspan class=\"CitationRef\"\u003e32\u003c/span\u003e].\u003c/p\u003e\n \u003cp\u003eOnce the classifiers are trained over a historical data stream, the selected classifiers are added to the classifier pool 1, 2 as per their place and they are used for layer 1,2 predictions. Layer 2 receives layer1 prediction results for reference plus it does its own prediction, these are given as a combined input to layer 3. Where a metaclassifier is giving a result for prediction. Here, it is Adaptive Random Forest selected as a metaclassifier since its higher precision. Once the component classifiers are added to the classifier pool at tier 1 and 2 respectively, we have tested an incoming sliding window of actual real time flight data, (already underwent data preprocessing and feature engineering as shown in the system architecture - refer Fig.\u0026nbsp;1, and the corresponding data passes through two tiers of ensemble, and then through a metaclassifier before it gives a final classification result for each flight record, whether it belongs to a class of delayed flights (class 1) or to the class of not-delayed flights(class 0). We have conducted the experiments on scikit multiflow and Anaconda Jupiter platforms, (details in table 1) and we have achieved remarkable progress as compared with the similar systems that we found in the literature. Figures\u0026nbsp;4 and 5 show the representative evaluation we have achieved for A-MERIT-C\u0026rsquo;s performance over a real time data stream of flights.\u003c/p\u003e\n\u003c/div\u003e"},{"header":"6. Results and Performance Analysis","content":"\u003cp\u003eWe have experimented with real time stream datasets in order to investigate the performance of component classifiers for their selection into ensemble. Table\u0026nbsp;4 shows performance analysis of component classifiers. As we are dealing with stream data, we have to found out a golden midpoint between classifiers ability to correctly classify an incoming data instance, predicting accurately in very less time as much as possible .So, the evaluation parameters for classifiers performance are kept as accuracy, kappa and time (Table\u0026nbsp;4). Na\u0026iuml;ve Bayes classifier although with average accuracy gain, it fits for the requirements of base layer [\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e33\u003c/span\u003e], hoeffding Tree classifier as mentioned in various data stream related research [\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e] [\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e27\u003c/span\u003e] is a good choice to be one of the classifiers of an ensemble. Also, KNN variant SAMKNN [ 28] is suitable as it has built in capacity of self-adjusting memory. As we have to design a heterogenous classifiers, the base and middle tier is composed of classifiers which are heterogenous enough, one of the main requirements of ensemble design in order to expect the better performance gain.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab4\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 4\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003ePerformance Analysis of component classifiers\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"4\"\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 \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eComponent Classifiers of Ensemble\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colspan=\"3\" nameend=\"c4\" namest=\"c2\"\u003e \u003cp\u003ePrequential Evaluation\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eAccuracy\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eKappa\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eTime(Sec)\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNa\u0026iuml;ve Bayes Classifier\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u003cb\u003e0.693\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.1513\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e\u003cb\u003e76.76\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSAMKNN - Self-Adjusting Memory KNN\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u003cb\u003e0.8013\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.5902\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e\u003cb\u003e34.18\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHoeffding Tree\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u003cb\u003e0.7767\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.5394\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e\u003cb\u003e11.27\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eExtremely Fast Decision Tree\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.7886\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.5668\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e467.93\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eThe design of multitiered ensemble relies on a metaclassifier at the final tier, so we need to choose a metaclassifier with optimized performance gain, studies from literature through the works of [\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e, \u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e, \u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e, \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e] have suggested the probable metaclassifiers as Logistic regression [\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e], SVM [\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e]and ARF[\u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e36\u003c/span\u003e]. We have continued our experimentation for a good choice of a metaclassifier. All these results are obtained at intermediate level of ensemble design, but they have played a vital role and laid a strong foundation in the final multitier ensemble design.\u003c/p\u003e \u003cp\u003eIn order to evaluate A-MERIT-S\u0026rsquo;s performance, we as researchers have reviewed a long list of relevant research articles but very few of them have really taken into consideration live real time data. Also, it was irrelevant if at all we could have compared the flight data analysis with that of a generic data stream analysis system. In keeping the same objective in mind, we found a similar research work in, where Souza et.al.[\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e] have worked with live real time flight data through various logistic regression models. After doing comparative performance analysis with that of [\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e], it is clear that our model A-MERIT-C has achieved considerable gain in evaluation parameters like f1, precision, recall. Also, accuracy of our model is at par with that of [\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e]. Table\u0026nbsp;5 shows the example results of this experimentation\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab5\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 5\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eA-MERIT-C performance evaluation\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"4\"\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 \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSr.\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eMetric\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eFLRDAR-24\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eA-MERIT-C\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eF1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e72.85\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e87.48\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003ePrecision\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e71.28\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e100\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eAccuracy\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e79.57\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e79.27\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eRecall\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e75.25\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e77.75\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eAUC\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e87.43\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e88.889\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\u003e \u003c/p\u003e \u003cp\u003eFigure\u0026nbsp;4 shows the comparative performance analysis where it is clear that model A-MERIT-C has achieved considerable gain in evaluation parameters like f1, precision, recall. Also, Accuracy of the model is at par with that of [\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e].Aviation stream data analysis is also taken further with other aviation datasets \u0026ndash; on time performance data \u0026ndash;as described below. A-MERIT-C with prequential evaluation is tested for the data. It has shown the remarkable performance (see Fig.\u0026nbsp;5 ) when compared with the benchmarked online active learning ensembles as shown in the table 6.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab6\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 6\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003ePerformance evaluation of \u0026ndash; AMERIT-C\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"4\"\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 \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eData\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eMethodology\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eMean Acc.(%)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eMean Time (S)\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"3\" rowspan=\"4\"\u003e \u003cp\u003e\u003cb\u003eairline-on-time performance data\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u003cb\u003eOnline AUE\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e77.06\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e7.26\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u003cb\u003eDWM\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e77.29\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e7.27\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u003cb\u003eACE\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e79.61\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e13.00\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u003cb\u003eA-MERIT-C\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e80.15\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e12.50\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"3\" rowspan=\"4\"\u003e \u003cp\u003e\u003cb\u003eBOT-US-DATA\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u003cb\u003eOnline AUE\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e99.23\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e30.86\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u003cb\u003eDWM\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e95.92\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e23.46\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u003cb\u003eACE\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e97.42\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e45.71\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u003cb\u003eA-MERIT-C\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e98.11\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e52.20\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\u003eIt is more obvious from the observations that all stream processing algorithms that too if they are online active learning ensembles, time taken for learning matters most, if accuracy is more. For airline -on time performance data its huge voluminous data approx. 120\u0026nbsp;million records with approx. 25\u0026ndash;30 features, the inherent veracity and volume has limited the accuracy evaluation. On the other hand, Bot-US-Data is balanced enough, so it has taken accuracy to highest level but only at the cost of mean time increased on an average.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eThis observation is very representative as working with real time data, since one has to find out a golden midpoint between time and accuracy. So, the models should be able to predict in as less time as possible and will be able to produce commendable progress in accuracy. It is also possible to evaluate the system for other performance parameters such as validation loss, MSE, prequential AUC, but the scope for the paper is kept limited only to the accuracy, precision and mean time taken approx. The research work is still continued for the extension.\u003c/p\u003e"},{"header":"7. Concluding remarks and discussions","content":"\u003cp\u003eThis research is very primitive attempt to address near real time stream data classification task w.r.t. aviation streams. There are some observations which will be definitely helpful for the novice researchers at least , who wish to work in the similar area.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e1. In case of live real time data, feature engineering and feature optimization play a very vital role in order to prepare data for analysis and to reduce dimensionality , which in turn further leads to minimum computational overhead for data analysis algorithms.\u003c/p\u003e\n\u003cp\u003e2. Using Stacking with three layers serves better as long as time required for analysis and computational efficiency is considered.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e3. Concept Drift Adaptation is widely taken care of and contributes in great extent when dealing with dynamic real time data stream. This is made possible via incrementally learning the algorithms in classifier pool through recent past historical data and always selecting the best performing classifiers to be the part of stacked ensemble learning.\u003c/p\u003e\n\u003cp\u003e4. Also, metaclassifiers like ARF used here are capable of dealing with drift adaptation. It is an added advantage of A-MERIT-C.\u003c/p\u003e\n\u003cp\u003e5. Keeping a learnt classifier always ready in a classifier pool leads to reducing the time required to train the classifiers drastically, as decision tress and random forests although yield very good performance , they have high demand for train time. In order to tackle with this issue , A-MERIT-C always keeps the pool of classifiers , ready to test an incoming data stream on the fly.\u003c/p\u003e"},{"header":"8. Constraints, limitations and Future scope ","content":"\u003cp\u003eIt\u0026rsquo;s a very true observation as working with the real time data stream analysis that , there is still scarcity of benchmarked platforms for computation. Although, real time stream analysis is need of the hour , not a plenty of standardised algorithms are being explored by research community ; (with very few exceptions). So, the desire of building a real time system concludes with actually building near-real time when facing challenges in data stream analysis[22,24] . As all real time systems are to be greatly improved with human intervention , same is the case with this research, if added with human intervention-intelligence , this can be of great help. This research lacks this aspect and can be taken further. One more aspect to be noted here is , this research does not take into consideration any weather delays, or catastrophic incidents in aviation stream analysis \u0026nbsp;just to keep the system complexity to the minimum level. This can be taken for further extension of the research. Further work can be inclusion of parallelization of some classifiers ---in GPU and other scalable computing environments (Scalable ensemble).\u003c/p\u003e"},{"header":"Declarations","content":"\u003ch2\u003eAuthor Contribution\u003c/h2\u003e\u003cp\u003eShailaja Jadhav , a corresponding author has conducted research , prepared manuscript .D. Kodavade has reviewed , corrected and guided the manuscript writing .\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n \u003cli\u003eSouza, V.M.A., dos Reis, D.M., Maletzke, A.G. et al. : Challenges in benchmarking stream learning algorithms with real-world data. Data Mining and Knowledge Discovery 34, 1805\u0026ndash;1858 (2020)\u003c/li\u003e\n \u003cli\u003eLeite, D., \u0026Scaron;krjanc, I. \u0026amp; Gomide, F. An overview on evolving systems and learning from stream data. Evolving Systems 11, 181\u0026ndash;198 (2020). https://doi.org/10.1007/s12530-020-09334-5\u003c/li\u003e\n \u003cli\u003eJicheng Shan , Hang Zhang, Weike Liu, and Qingbao Liu- \u0026ldquo;Online Active Learning Ensemble Framework for Drifted Data Streams\u0026rdquo; - IEEE Transactions on Neural Networks and Learning systems.vol \u0026ndash; 6 Issue 1- 2162-237X \u0026copy; 2018\u003c/li\u003e\n \u003cli\u003eQi Wang, ZhiHao Luo, JinCai Huang, YangHe Feng, and Zhong Liu \u0026ldquo;A Novel Ensemble Method for Imbalanced Data Learning: Bagging of Extrapolation-SMOTE SVM \u0026rdquo;\u0026ndash; Hindawi Computational Intelligence and Neuroscience Volume 2017\u003c/li\u003e\n \u003cli\u003eHeitor Murilo Gomes, Jean Paul Barddal, and Fabr\u0026acute;icio enembreck , A Survey on Ensemble Learning for Data Stream Classification\u0026rdquo;, ACM Computing Surveys, Vol. 50, No. 2, Article 23, Publication date: March 2017\u003c/li\u003e\n \u003cli\u003eLeszek Rutkowski, Maciej Jaworski, Piotr Duda, \u0026ldquo;Stream Data Mining: Algorithms and Their Probabilistic Properties \u0026rdquo; ISBN 978-3-030-13961-2 eBook\u003c/li\u003e\n \u003cli\u003eTingting Zhai\u0026ndash; \u0026ldquo; Classification of high-dimensional evolving data streams via a resource-efficient online ensemble\u0026rdquo; \u0026ndash;Springer - Data Mining and Knowledge Discovery September 2017, Volume 31, Issue 5, pp 1242\u0026ndash;1265\u003c/li\u003e\n \u003cli\u003eDariusz Brzezinski, Jerzy Stefanowski, \u0026ldquo;Combining block-based and online methods in learning ensembles from concept drifting data streams\u0026rdquo;, Information Sciences, Volume 265, 2014, Pages 50-67\u003c/li\u003e\n \u003cli\u003eLeite, D., \u0026Scaron;krjanc, I. \u0026amp; Gomide, F. An overview on evolving systems and learning from stream data. Evolving Systems 11, 181\u0026ndash;198 (2020).\u003c/li\u003e\n \u003cli\u003eHeitor Murilo Gomes, Jean Paul Barddal, and Fabr\u0026acute;icio enembreck , \u0026ldquo; A Survey on Ensemble Learning for Data Stream Classification\u0026rdquo;, ACM Computing Surveys, Vol. 50, No. 2, Article 23, Publication date: March 2017\u003c/li\u003e\n \u003cli\u003eBartosz Krawczyk, Leandro L. Minku, Jo\u0026atilde;o Gama, Jerzy Stefanowski, MichałWoźniak, \u0026ldquo;Online Active Learning Ensemble Framework for Drifted Data Streams\u0026rdquo;, Information Fusion, Volume 37,2017, Pages 132-156\u003c/li\u003e\n \u003cli\u003eMinku, Leandro \u0026amp; Yao, Xin. DDD: A New Ensemble Approach for Dealing with Concept Drift. Knowledge and Data Engineering, IEEE Transactions on. 24. 619 \u0026ndash; 633, (2012).\u003c/li\u003e\n \u003cli\u003eBrzezinski, D., Stefanowski, J. , Prequential AUC: properties of the area under the ROC curve for data streams with concept drift. Knowledge Information Systems 52, 531\u0026ndash;562 (2017).\u003c/li\u003e\n \u003cli\u003eElwell R. and Polikar R., \u0026ldquo;Incremental Learning of Concept Drift in Nonstationary Environments\u0026ldquo;, IEEE Trans. on Neural Networks, vol. 22, 1517-1531, 2011.\u003c/li\u003e\n \u003cli\u003eBertini Junior, J.R., do Carmo Nicoletti,M.: An iterative boosting-based ensemble for streaming data classification. Inf. Fusion 45, 66\u0026ndash;78 (2019)\u003c/li\u003e\n \u003cli\u003eKrawczyk, B., Cano, A.: Online ensemble learning with abstaining classifiers for drifting and noisy data streams. Appl. Soft Comput. 68, 677\u0026ndash;692 (2018)\u003c/li\u003e\n \u003cli\u003eScikit-multiflow :Montiel, J., Read, J., Bifet, A., \u0026amp; Abdessalem, T. (2018). Scikit-multiflow: A multi-output streaming framework. The Journal of Machine Learning Research, 19(72):1\u0026minus;5.\u003c/li\u003e\n \u003cli\u003eZ. Yu et al., \u0026quot;Hybrid Incremental Ensemble Learning for Noisy Real-World Data Classification,\u0026quot; in IEEE Transactions on Cybernetics, vol. 49, no. 2, pp. 403-416, Feb. 2019, doi: 10.1109/TCYB.2017.2774266\u003c/li\u003e\n \u003cli\u003eAviationStack: https://aviationstack.com\u003c/li\u003e\n \u003cli\u003eKluyver, T. et al., 2016. Jupyter Notebooks \u0026ndash; a publishing format for reproducible computational workflows. In F. Loizides \u0026amp; B. Schmidt, eds. Positioning and Power in Academic Publishing: Players, Agents and Agendas. pp. 87\u0026ndash;90.\u003c/li\u003e\n \u003cli\u003eAljubairy, A., Zhang, W.E., Shemshadi, A. et al. A system for effectively predicting flight delays based on IoT data. Computing 102, 2025\u0026ndash;2048 (2020).\u003c/li\u003e\n \u003cli\u003eSouza, V.M.A., dos Reis, D.M., Maletzke, A.G. et al. Challenges in benchmarking stream learning algorithms with real-world data. Data Mining and Knowledge Discovery, vol. 34, 1805\u0026ndash;1858 (2020).\u003c/li\u003e\n \u003cli\u003eQi Wang, ZhiHao Luo, JinCai Huang, YangHe Feng, and Zhong Liu \u0026ldquo;A Novel Ensemble Method for Imbalanced Data Learning: Bagging of Extrapolation-SMOTE SVM \u0026rdquo;\u0026ndash; Hindawi Computational Intelligence and Neuroscience Volume 2017, Article ID 1827016\u003c/li\u003e\n \u003cli\u003eAlonso-Betanzos, Amparo ,Pavel, Rokas, Kurasova. Prediction of Flight Time Deviation for Lithuanian Airports Using Supervised Machine Learning Model, Computational Intelligence and Neuroscience, Hindawi, SP-8878681,VL - 2020\u003c/li\u003e\n \u003cli\u003eGomes, H.M., Bifet, A., Read, J. et al. Adaptive random forests for evolving data stream classification. Mach Learn 106, 1469\u0026ndash;1495 (2017).\u003c/li\u003e\n \u003cli\u003eKolter, J. Z., \u0026amp; Maloof, M. A. Dynamic Weighted Majority: An ensemble method for drifting concepts. Journal of Machine Learning Research 8:2755\u0026ndash;2790. (2007).\u003c/li\u003e\n \u003cli\u003eGonzalez Hidalgo, Juan \u0026amp; Maciel, Bruno Iran Ferreira \u0026amp; Barros, Roberto. Experimenting with prequential variations for data stream learning evaluation. Computational Intelligence. 35. 670-692, (2019).\u003c/li\u003e\n \u003cli\u003eH.-W. Hu, Y.-L. Chen, K. Tang, A dynamic discretization approach for constructing decision trees with a continuous label, IEEE Trans. Knowl. Data Eng. 21 (11) (2009) 1505\u0026ndash;1514.\u003c/li\u003e\n \u003cli\u003eV. Losing, B. Hammer and H. Wersing, \u0026quot;KNN Classifier with Self Adjusting Memory for Heterogeneous Concept Drift,\u0026quot; 2016 IEEE 16th International Conference on Data Mining (ICDM), Barcelona, Spain, 2016\u003c/li\u003e\n \u003cli\u003eD. Brzezinski, J. Stefanowski, R. Susmaga and I. Szczech, \u0026quot;On the Dynamics of Classification Measures for Imbalanced and Streaming Data,\u0026quot; in \u003cem\u003eIEEE Transactions on Neural Networks and Learning Systems\u003c/em\u003e, vol. 31, no. 8, pp. 2868-2878, Aug. 2020.\u003c/li\u003e\n \u003cli\u003eMatthias Sch\u0026auml;fer, Martin Strohmeier, Vincent Lenders, Ivan Martinovic and Matthias Wilhelm. \u0026quot;Bringing Up OpenSky: A Large-scale ADS-B Sensor Network for Research\u0026quot;. In Proceedings of the 13th IEEE/ACM International Symposium on Information Processing in Sensor Networks (IPSN), pages 83-94, April 2014.\u003c/li\u003e\n \u003cli\u003eAirline On-time Performance Data : https://www.kaggle.com/datasets/bulter22/airline-data\u003c/li\u003e\n \u003cli\u003eGomes, H.M., Bifet, A., Read, J. et al. Adaptive random forests for evolving data stream classification. Mach Learn 106, 1469\u0026ndash;1495 (2017).\u003c/li\u003e\n \u003cli\u003eY. Yang, G.I. Webb, Discretization for Naive\u0026ndash;Bayes learning: managing discretization bias and variance, Mach. Learn. 74 (1) (2009) 39\u0026ndash;74.\u003c/li\u003e\n \u003cli\u003eJ. Kunnen, M. Duchateau, Z. Van Veldhoven and J. Vanthienen, Benchmarking Stacking Against Other Heterogeneous Ensembles in Telecom Churn Prediction, 2020 IEEE Symposium Series on Computational Intelligence\u003c/li\u003e\n \u003cli\u003eLiu, N., Gao, H., Zhao, Z. et al. A stacked generalization ensemble model for optimization and prediction of the gas well rate of penetration: a case study in Xinjiang. J Petrol Explor Prod Technol \u003cstrong\u003e12\u003c/strong\u003e, 1595\u0026ndash;1608 (2022).\u003c/li\u003e\n \u003cli\u003eGomes, H.M., Bifet, A., Read, J. et al. Adaptive random forests for evolving data stream classification. Mach Learn 106, 1469\u0026ndash;1495 (2017).\u003c/li\u003e\n \u003cli\u003eJadhav, Shailaja \u0026amp; Kodavade, D.. (2023). Enhancing Flight Delay Prediction through Feature Engineering in Machine Learning Classifiers: A Real Time Data Streams Case Study. International Journal on Recent and Innovation Trends in Computing and Communication. 11. 212-218. 10.17762/ijritcc.v11i2s.6064.\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"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":"adaptive ensemble, prequential analysis, stream data mining, streaming data analytics ","lastPublishedDoi":"10.21203/rs.3.rs-5847010/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-5847010/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eRecently, there has been a lot of study focused on large-scale data analysis. Data typically arrives continually in the form of data streams in numerous applications of today’s data intensive world. The analytics engines designed to deal with streaming data need to be capable of acting on data -in- motion . Since the conventional techniques for data mining and machine learning were created for static datasets, data streams present unique difficulties. They are unable to effectively analyse data that is expanding at a quick pace and are less suitable to take into account the representative qualities of data streams. The authors through this research viz. \u003cu\u003eA-MERIT-C\u003c/u\u003e(\u003cu\u003e\u003cstrong\u003eA\u003c/strong\u003e\u003c/u\u003edaptive \u003cu\u003e\u003cstrong\u003eM\u003c/strong\u003e\u003c/u\u003eultitiered \u003cu\u003e\u003cstrong\u003eE\u003c/strong\u003e\u003c/u\u003ensemble for \u003cu\u003e\u003cstrong\u003eR\u003c/strong\u003e\u003c/u\u003eeal (\u003cu\u003e\u003cstrong\u003eI\u003c/strong\u003e\u003c/u\u003entegrated with feature engineering) \u003cu\u003e\u003cstrong\u003eT\u003c/strong\u003e\u003c/u\u003eime \u003cu\u003e\u003cstrong\u003eC\u003c/strong\u003e\u003c/u\u003elassification), have presented an active learning data stream analysis model through Ensemble Learning Framework ,which is able to learn concepts in near real time streaming data environment. The design of the framework is adaptive to address the dynamism associated with real time data through the classifier pool (i.e. it chooses the best performing classifiers). This design of Adaptive Multitiered Ensemble is composed of prequentially evaluated classifiers rather than traditional hold out evaluation. The distinguishing characteristics of A-MERIT-C give considerable improvements in terms of f1 (87.48), recall (77.79), accuracy (79.27) AUC (88.89) for streaming data analytics as compared to that of similar real time airline FLTRADAR 24 models ; on the other hand, it is able to address the limitations of existing algorithms such as concept evolution , feature drift through incremental learning and feedback .\u003c/p\u003e","manuscriptTitle":"Adaptive Multilayer Ensemble based Real Time Stream data Classification : Architectural Perspectives and Performance","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-02-10 11:33:11","doi":"10.21203/rs.3.rs-5847010/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":"0e184487-957c-4df3-ae22-46907e78729d","owner":[],"postedDate":"February 10th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2025-02-15T11:38:40+00:00","versionOfRecord":[],"versionCreatedAt":"2025-02-10 11:33:11","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-5847010","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-5847010","identity":"rs-5847010","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}