Author(s): Tahmida Naher Chowdhury; Rajat Nag; Md Arman Habib; Md Salauddin
Linked Author(s):
Keywords: Groundwater Contamination Machine learning Random Forest Support Vector Machines Water Quality Index
Abstract: Groundwater is a crucial source of freshwater for drinking, irrigation, and industrial use. However, its quality is reportedly at risk due to contamination from both natural and anthropogenic sources. Consequently, analysing and monitoring groundwater quality is critically essential for safeguarding human health, maintaining sustainable agricultural and industrial practices, protecting ecosystems, and successfully managing water supplies for both present and future needs. While reliable, traditional analytical methods for groundwater quality assessment face challenges in handling large and complex datasets and offering real-time or predictive insights, integrating machine learning (ML) techniques presents a promising alternative by leveraging data-driven approaches to enhance the accuracy and efficiency of groundwater quality predictions. This research presents the performance of five advanced ML algorithms (both kernel-based and decision support trees), including Support Vector Machines (SVM), Artificial Neural Networks (ANNs), Random Forest (RF), Gradient Boosting (GBoost), and K-nearest neighbours (KNNs) in assessing the groundwater quality status as well as predicting the groundwater quality index in Ireland. A variety of error measures, such as Mean Squared Error (MSE), Root Mean Squared Error (RMSE), and the Coefficient of Determination (R2), were calculated to evaluate the performance of ML models in groundwater quality assessment and prediction. The predictive ML models as developed within this work would improve understanding of the complex interactions by evaluating the impact of climate change on groundwater resources as well as help us to predict groundwater quality parameters and contamination levels and forecast future trends.
Year: 2025