The main objective of this study is to predict and monitor groundwater quality through the use of modern Machine Learning (ML) techniques. By employing ML techniques, the research effectively evaluates groundwater quality to forecast its future trends. Five machine learning models Logistic Regression (LR), Support Vector Machine (SVM), Random Forest (RF), Adaptive Boosting (AdaBoost), Extreme Gradient and Boosting (XGBoost) were used here to predict the water quality by assessing the physical and chemical parameters such as electrical conductivity (EC), hydrogen ion (pH) concentration, total dissolved solids (TDS), chemical parameters such as, sodium (Na+), magnesium (Mg2+), calcium (Ca2+), potassium (K+), bicarbonates (HCO3-), fluoride (F-), sulphate (SO42-), chloride (Cl-), and nitrate (NO3-) in 94 dug and bore wells from the semi-arid river basin (Arjunanadi) in Tamil Nadu, India. The pH of the samples is alkaline nature. Gibb's diagram suggested the rock-water dominance and minor influence of evaporation and crystallization on the hydrochemistry. From water quality index, 599.75 km2 (53%) of area has a good quality and 536.75 km2 (47%) of area has poor water quality. Water Quality Index values (WQI) of water quality formed baseline data for the prediction models as a dependent variable, and the physicochemical parameters were used as independent variables. The model efficacies were assessed using statistical error such as Relative Squared Residual (RSR) error, Nash-Sutcliffe efficiency (NSE), Mean Absolute Percentage Error (MAPE), Coefficient of determination (R2) and final accuracy. In this study, the LR model provided the minimal error (RSR = 0.22, NSE = 0.95, MAPE = 1.3) with an accuracy of 95% in predicting the water quality. The performance of the ML models is in the sequence of SVM > Adaboost > XGBoost > RF. This study helps the lawmakers and administrators for creating awareness on modern techniques for predicting and monitoring groundwater quality on the general public and supporting to achieve the sustainable development goals 3 and 6 for clean and healthy community.
Read full abstract