Advancing groundwater quality assessment in Siliguri City of India through the RMS-WQI model incorporating the data-driven approaches

• First study to apply a data-driven approach for monitoring GWQ in eastern India. • A novel RMS-WQI model utilized for assessing groundwater suitability. • GWQ ranged from fair to good across the study area . • The XGB model with OPT optimizer provided accurate GWQ. • Model showed high sensitivity and low uncertainty for assessing GWQ. As a vital renewable resource, groundwater is significant and safeguarded source of drinking water for city population. Therefore, continuous assessment of the groundwater quality (GWQ) is required to maintain the good status of the GWQ. The current research focused on assessing the GWQ of Siliguri City, eastern India, utilizing the root mean squared water quality index (RMS-WQI) model. Furthermore, the research incorporated the eXtreme Gradient Boosting (XGB) algorithm with Optuna (OPT) (XGB-OPT) hyperparameter optimization technique to enhance the accuracy of predictions and reduce potential errors within the RMS-WQI model. The computed WQI scores ranged from 66.4 to 94.0, with most sampling locations falling within the ‘Good’ to ‘Fair’ classification. Moreover, the XGB-OPT model showed high accuracy in both the training and testing period (Training RMSE = 3.27, Testing RMSE = 1.50; Training MSE = 9.74, Testing MSE = 2.25; Training MAE = 3.01, Testing MAE = 0.69; and Training PABE = 3.41 and Testing PABE = 0.79) in terms of WQI score prediction. Additionally, the developed model showed high sensitivity (R 2 = 0.94) and low uncertainty (< 1 %) in terms of WQI score prediction with the least prediction error at each sampling sites (PREI = 0.096; NSE = 0.816; MEF = 0.815). These results collectively demonstrated the robustness and reliability of the RMS-WQI model in assessing the GWQ. The findings highlighted the effectiveness of the RMS-WQI model in the GWQ assessment. The research outcome presents critical insights for the stakeholders, including policymakers, environmental regulators, and scientific researchers, facilitating more informed and strategic decision-making for sustainable groundwater management. By demonstrating the reliability and precision of the RMS-WQI model, this research contributed to improve the existing GWQ monitoring approach, which could ultimately support long-term water security scheme and provide safeguard to public health in Siliguri City.