Machine Learning Models for the Prediction of Water Quality Index Using Physicochemical Characteristics of Groundwater

¹ Department of Engineering and Technology, College of Engineering, University of Technology and Applied Sciences, Muscat, Oman., Oman.
² Department of Civil Engineering, Dayananda Sagar College of Engineering, Bengaluru - 560 111, Karnataka, India
³ Civil Engineering Department, Al-Qalam University, Kirkuk, Iraq
⁴ Civil Engineering Department, Tishk International University, Sulaimani, Iraq. Email: This email address is being protected from spambots. You need JavaScript enabled to view it. , ORCID: 0000-0002-8225-1721
⁵ Department of Civil Engineering, University College of Engineering, Thirukkuvalai, Nagapattinam, Tamil Nadu, India

^* Corresponding author: This email address is being protected from spambots. You need JavaScript enabled to view it.

Abstract

This research examines the capacity of machine learning (ML) methods to predict the Water Quality Index (WQI) of groundwater from a set of physicochemical parameters. A dataset of 409 groundwater samples was examined, which includes pH, EC (electrical conductivity), TH (total hardness), Ca (calcium), Mg (magnesium), Na (sodium), K (potassium), HCO₃ (bicarbonate), Cl (chloride), SO₄ (sulfate), and NO₃ (nitrate). The WQI was calculated and employed as the target variable in this analysis. Four regression models (Decision Tree Regressor, Random Forest Regressor, Gradient Boosting Regressor, XGBoost Regressor) were evaluated using statistical metrics. The Gradient Boosting Regressor demonstrated superior predictive capability among the four tested models, with an R² = 0.8982, MAE = 9.2954, and MSE = 1361.6065. XGBoost and Random Forest models also have high predictive capacity; however, the Decision Tree model has relatively low accuracy compared to the other three models. The results suggest that ML algorithms can identify non-linear associations between groundwater quality parameters and WQI. This research shows that data-driven approaches can provide rapid and reliable predictions of groundwater quality, which may help reduce the need for extensive laboratory testing and promote better utilisation of water resources through more effective and efficient water resource management.