Water Quality Classification Using Machine Learning

Abstract

The quality of water has been significantly influenced by many other pollutants over the past few years. It has the direct impact on human health and environment. The WQI works as an indicator of water management, efficiency. Knowing the quality of water and even how to model the quality in prediction benefits the war against water pollution. The objective of the study is to establish a reliable prediction model for river water quality, that is able to identify the index value is based on the river water quality standards. In this project, the project Inspect and compare the performance of many classification models and algorithms to find which attributes were prominent in classifying river water quality. A total of eleven sampling stations, spread across different points on the River flowing through Kerala and Tamil Nadu, have been chosen for the data collection. The water quality index is measured by 7 different environmental factors affecting the quality of water, including the dissolved oxygen level, temperature, pH, hardness, chloride etc. Supervised Machine learning algorithms such as logistic regression, support vector regressor have been used to develop a model for predicting water quality. A classification model was developed using SVM classifiers, SVM, to classify water quality index. Logistic regressor efficiently predicts water Quality index, SVM classifier classifies water Quality index with an accuracy of 83%. The built models presented favourable results regarding water quality index predictions and classification.