Toward Smart Water Quality Monitoring:Potability Prediction using Supervised Machine Learning

Access to safe drinking water remains a cornerstone of public health and human development, yet ensuring its quality poses persistent challenges in many regions. This study aims to leverage advanced machine learning (ML) models to predict drinking water quality from a public health nursing perspective. Ensuring the safety of drinking water is critical for public health, environmental sustainability, and the achievement of the United Nations Sustainable Development Goals. Traditional approaches to water quality assessment are often labor-intensive, time-consuming, and limited in scalability, rendering them inadequate for real-time or large-scale monitoring. In this research, we evaluate the predictive capabilities of several ML algorithms including Decision Tree (with hyperparameter tuning), Multilayer Perceptron (MLP), Recurrent Neural Network (RNN), Deep Neural Network (DNN), Support Vector Machine (SVM), XGBoost, and AdaBoost for the classification of drinking water potability using physicochemical parameters. Experimental findings indicate that the RNN model achieves the highest classification accuracy at 90%, followed by SVM (89%), MLP (88%), and DNN (85%). These results underscore the potential of advanced ML methodologies in developing accurate, efficient, and scalable water quality monitoring systems that can significantly contribute to public health protection and policy implementation.