A Machine Learning Based Method for Improving the Performance of Water Quality Prediction

The assessment of water resources plays an important role in human activities and daily production. Not only does this task help direct users to have a water source of guaranteed quality and meet their needs, but it also helps managers take timely actions to keep the water source clean and usable. In recent times, machine learning algorithms have been used a lot in the literature to evaluate water quality and many of them have achieved impressive accuracy. Having said that, these sophisticated algorithms nevertheless fail to meet expectations in a few case studies. Enhancing the performance of the ML-based approach to the water quality evaluation problem using a real-world dataset is the primary objective of this study. To further reduce the amount of input parameters for the models, we think about applying the feature selection method to pick the important characteristics for the ML models. This has important practical implications, allowing for saving time and costs in assessing water quality.