Apprising the Effectiveness in Predicting the Quality of Indian Rivers Using Two-layered Ensemble Learning Approach

Water contamination is indeed a major issue in Indian rivers, and the origins of contamination are diverse, including various human performances such as industrial wastes, domestic manure, farming runoff, and religious follows. The inferences of water pollution are severe and far-reaching, affecting the health of living organisms and the environment. Water pollution can cause various health problems, including intestinal illnesses, skin infections, neurological illnesses, etc. These health issues can be particularly challenging in areas where people rely on rivers for drinking, bathing, and other domestic uses. Estimating is an approach for watershed controlling that identifies common restrictions related to conventional outdated methods for calculating the quality of water. Appraising the effectiveness in predicting the magnitude of water contamination and for developing effective strategies to manage and prevent water pollution Machine learning (ML) algorithms with multilayer ensemble learning algorithms are broadly and successfully used for decision-making issues such as more inconsistent conditions of rivers in India. This research proposes the different ML algorithms for ensemble learning such as linear regression (LR), support_vector_regression (SVR), random_forest_regression (RFR), and bayesian_network (BN) as stacking model contains two-layered ensemble learning technique. The output predictions of the four commonly utilized individual machine learning models mentioned earlier are utilized as input features for a higher-level stacked LR model. This stacked LR model produces the final prediction with an impressive accuracy of 99.95%. By implementing two layered ensemble learning approaches to various rivers in India, to show the accurate predictions results associated with all four first layer ML models.