Comparison of Water Quality Classification Models using Machine Learning
Neha Radhakrishnan, Anju S. Pillai
Abstract
Water resources are often polluted by human intervention. Water pollution can be defined in terms of its quality which is determined by various features like pH, turbidity, electrical conductivity dissolved oxygen (DO), nitrate, temperature and biochemical oxygen demand (BOD). This paper presents a comparison of water quality classification models employing machine learning algorithms viz., SVM, Decision Tree and Naïve Bayes. The features considered for determining the water quality are: pH, DO, BOD and electrical conductivity. The classification models are trained based on the weighted arithmetic water quality index (WAWQI) calculated. After assessing the obtained results, the decision tree algorithm was found to be a better classification model with an accuracy of 98.50%.