Accurate Forecast of Water Quality Index for Cholera Diseases using Two-Layered Stacked Machine Learning Algorithms

Enhancing accuracy and reducing the error prediction of water quality index regarding the quality criteria for V Cholera diseases is the main aspect of day-to-day life. The implementation of this work utilises the Two-Layered Stacked Ensemble Model, which consists of four base models: Multiple Linear Regression (MLR), Partial Least Square Regression (PLSR), Sparse Partial Least Square Regression (SPLSR), and Random Forest Regression (RFR). These models are used in the first layer, while the second layer is composed of the stacked Ensemble RFR (ERFR). The quality parameters of water are underwent preprocessing and were utilised to compute Water Quality Index (WQI) through the application of the weighted arithmetic mean technique. Accurate prediction of water quality utilized from four base algorithms and stacked Ensemble RFR. Performance metrics like mean squared error, mean absolute error, root mean squared error, and R squared values are used for appraising overall performances of the proposed model. The degree of accuracy obtained for the suggested model is the basis for their evaluation. The accuracy in predicting the WQI from the dataset using MLR (97.15%), PLSR (96.99%), SPLSR (95.01%), RFR (93.59%) and Stacked Ensemble RFR (99.12%) is obtained. Numerous principles and two-layered ensemble with stacked models were utilised to provide V Cholera WQI predictions. The application of ensemble methods in predicting water quality and identifying pollution sources has achieved promising results from the stacked model with an accuracy of 99.12%, hence the Stacked Ensemble RFR is significantly better than the other base Regression algorithms.