Comparative Analysis of Machine Learning Algorithms for Water Quality Index Prediction

Abstract

Access to clean and safe water is essential for environmental sustainability and economic progress. The Water Quality Index (WQI) serves as an important metric, synthesizing complex water quality data into a single value to assess its suitability for various uses, including drinking, agriculture, and aquaculture. In the recent years, Machine Learning (ML) algorithms have revolutionized WQI prediction, offering efficient and scalable solutions to analyse complex datasets. A comparative study of the performance of multiple ML algorithms—Naïve Bayes, XGBoost, K-Nearest Neighbors, Decision Tree, AdaBoost, and Logistic Regression for WQI prediction is explored in this paper. Experimental results demonstrate that Naïve Bayes outperforms other classifiers, highlighting its suitability for handling probabilistic relationships in water quality data. The findings underscore the growing significance of ML to enhance real-time water quality monitoring and sustainable resource management, particularly for small-scale water systems. By leveraging ML, this research provides a foundation for more accurate and reliable water quality assessments.