فهرست نشریات علمی دسترسی آزاد ایران

Water quality is a critical determinant of ecological and public health, making its regular assessment essential for sustainable development. This study aims to estimate the Water Quality Index (WQI) using multiple water parameters—pH, temperature, dissolved oxygen (DO), conductivity, faecal coliform, and nitrate-nitrite nitrogen. The dataset, sourced from Kaggle, comprises water samples collected across 18 Indian states. A weighted arithmetic WQI approach is employed to compute the index values. To forecast WQI, four regression models, linear regression, decision tree, random forest, and gradient boosting, are applied. Model performance is evaluated using the coefficient of determination (R²). Among all models, gradient boosting achieved the highest prediction accuracy, with an R² value of 0.94, significantly outperforming the others. The results highlight the effectiveness of machine learning in modelling complex environmental parameters and forecasting water quality. This study demonstrates that data-driven approaches can support timely decision-making for water resource management and public health interventions.