Assessment of groundwater quality using water quality index, multivariate statistical analysis and machine learning techniques in the vicinity of an open dumping yard

The groundwater around the open dumping yard is very susceptible to pollution due to infiltration of landfill leachate, which has higher concentrations of harmful organic and inorganic substances. Hence, it is imperative to regularly examine the groundwater quality. 18 sample wells were identified near the open dumping site in Saduperi village, Vellore district, Tamil Nadu, India. Overall, 216 groundwater samples were taken for a year (2021–2022). Cluster analysis and principal component analysis are multivariate statistical studies that were used to discover the primary variables that influence groundwater quality. In addition, the water quality index (WQI) was computed with the arithmetic method and machine learning techniques (ML). ML techniques such as decision tree regression, linear regression, random forest regression and ridge regression (RR) have been used to predict WQI. More than 50% of the total wells around the dumping site were observed to have physiochemical parameters exceeding the permissible limit. The findings obtained using machine learning approaches reveal that the RR model outperforms other ML models, attaining the highest validation precision of R² = 0.999, mean absolute percentage error  = 3.195, root mean square error = 0.406, and mean absolute error = 1.149. The study findings illustrate the robustness of both arithmetic WQI and ML-WQI to analyse groundwater quality. Therefore, this study assists policy decision-makers in forecasting the groundwater quality surrounding the landfill location by the use of a precise ML model.