摘要
Proper evaluation of river water quality is the key to the sustenance of the eco-system, sustainable use of resources, and the preservation of human health.In the current work, we will present an artificial intelligence-based algorithm that will predict the Water quality index (WQI) of Indian rivers using machine learning algorithms, namely, Linear Regression, Random Forest, and eXtreme Gradient Boosting (XGBoost).The data set has integrated a large amount of physicochemical and weathering variables, i.e., biochemical oxygen demand (BOD), chemical oxygen demand (COD), dissolved oxygen (DO), pH, nitrates, temperature, and rainfall.Experimental analysis shows that XGBoost yields the optimum predictive results (R 2 = 0.89, RMSE = 5.1, MAE = 3.7), which is greater than baseline and ensemble choices.In order to improve the interpretability, we use SHAP (SHapley Additive exPlanations) analysis that selects BOD, COD, and DO as the most significant parameters that influence the prediction of WQI.Moreover, case studies on urban-rural river reaches show that the model is able to model seasonal changes, pollu-tion peaks and festival-related unique features.The results highlight the value of explainable AI in aiding the active surveillance of rivers to facilitate early alerts and make decisions regarding policy changes.Future directions that involve incorporation of real time sensor information, prediction of wastewater pollution, and climate resilience modeling have also been outlined in the study, thus playing a role in sustainable water governance and SDG 6 (Clean Water and Sanitation).