Predicting Water Potability Using Machine Learning: A Comparative Analysis of Classification Algorithms

Water scarcity is a growing concern in Malaysia and worldwide, necessitating innovative solutions for sustainable water management. This research investigates the application of various machine learning algorithms to predict the potability of water based on 20 chemical parameters. The study employs Logistic Regression, Naive Bayes, Decision Tree, Random Forest, K-Nearest Neighbors, Support Vector Machines, Extreme Gradient Boosting, and Multi-layer Perceptron to classify water samples as potable or non-potable. A comprehensive methodology is followed, including data preprocessing steps such as handling missing values, addressing data imbalance with SMOTE, and normalizing the dataset. The models are evaluated using metrics like F1-score and log loss to determine their accuracy and reliability. Results indicate that the XGBoost algorithm outperforms others with the highest F1-score of 0.9798 and the lowest log loss of 0.7150, demonstrating its efficacy in water quality prediction. This study highlights the potential of machine learning techniques in enhancing water quality monitoring systems, contributing to better water resource management and public health protection.