Real-Time Multi-Class Classification of Water Quality Using MLP and Ensemble Learning

The major goal of water management planning and the iterative evaluation of operational policies and procedures is to ensure that good water quality is always maintained. Effective water monitoring requires examining many water samples, which is a time-consuming and labor-intensive process that takes a lot of effort. This paper aims to evaluate the quality of drinking water samples with high accuracy by using multi-class classification models: multilayer perceptron (MLP) and ensemble learning. Real datasets with different sizes that include the essential water quality parameters have been used to train and test the developed models. The results showed the effectiveness of the developed models in detecting water contamination with high accuracy in both datasets used. The results demonstrate that bagging ensemble learning outperforms the multilayer perceptron with an overall accuracy of $$94\%$$ for station-A and $$92\%$$ for station-B compared to MLP, which shows an overall accuracy of $$89\%$$ for station-A and $$87\%$$ for station-B.