Prediction of Water Quality Using Artificial Intelligence (AI) and Statistical Approach

Water is a valuable resource on the earth. To be consumed safely, it must be treated. But the increasing pollution of the reserves makes this operation more and more delicate. Accordingly, water quality modeling and prediction have become critical in tackling pollution. The goal of this study is to use an advanced artificial intelligence (AI) system to estimate the water quality index (WQI) for the surface. We aim through this research to generate a non-supervised machine learning Model that can predict the value of target variable inferred from available data (training data), as a soft computing technique: Decision Trees for Classification was used to predict the potability of water. It provides a good performance in this study. A big dataset, collected monthly over a period of 5 years, was employed in the modeling process. The data were splited into sets, one for model building (training dataset) and the other for model validation (testing dataset) to avoid over-learn. Statistical assessment indicators and visual evaluation Metrics estimated the prediction capacity of the developed model. The outcome demonstrates that DTC was opportune for the forecast of the potability indice and would be functional for water purification organizations in the management of divers water supply systems (R-2 = 72%, RMSE = 0.41). The models developed in this study could serve as a foundation for decision-making that will aid in the maintenance and improvement of water supply system management.