ANALYSIS OF MACHINE LEARNING MODELS FOR EFFICIENT WATER QUALITY PREDICTION

Abstract

Water plays a vital role in every aspect of human life, including the metabolism of organisms, industrial manufacturing of goods, and so on. Since water is essential to humanity and is used to improve our way of life, it will be a front-line challenge to humanity if it is heavily contaminated by their activities. In this study, a system that determines an efficient Machine Learning (ML) Model for better water quality prediction is proposed. The performance of the proposed model is evaluated in two different directions: the classification accuracy and the model accuracy in terms of quality (precision) and quantity (recall) of the prediction output. It is identified that the classification accuracy of the Random Forest algorithm appears to be the best model for water quality prediction on the obtained dataset. Random Forest outperforms other algorithms with the best accuracy score of 0.87 as against the XGBoost with 0.86. The model will be useful for treatment plants by automating and training the procedure of determining the quality of the water sample since the water quality of the water sample must be identified to determine the right amount of chemicals to be introduced for the water sample to be potable. In the near future, this work is aimed to be deployed with Flask to provide an interactive interface for users with a non-technical background to use the standard parameters in assessing the quality of water being used for daily activities.