An Intelligent Approach for Investigating Water Quality Using Machine Learning Models

Abstract

Potable or Drinking Water quality is a worldwide concern. The growth of habitat fragmentation has an effect on water quality, and any contamination, whether physical or chemical, lowers the dependability of the receiving water body. This research forecasts the safest drinking water premised on certain factors such as Hardness, Turbidity, pH factor, Presence of sulfate, and many more. A machine learning centered strategy has been approached by this study to carry out the automated water quality monitoring. This investigation uses various boosting and bagging ensemble classification approaches. Random Forest and Extra Tree are the bagging classification methods whereas; Gradient Boost, Extra Gradient Boost, Light Gradient Boost, CatBoost, and AdaBoost methods are implemented for boosting ensemble classification. An exhaustive comparison among these ensemble models has been conducted for inferring the best predictive model. After hyper-tuning each model, it was discovered that the Random Forest provides the best predicting accuracy of 80.95%.