Abstract:

The Water quality has deteriorated significantly over the last few decades owing to contamination and other factors. As a result, a model capable of making accurate estimates regarding water quality is required. Keeping track of the treated water outflow is critical for the stability and conservation of the environment. Moreover, inadequate sanitary facilities and a lack of knowledge contribute significantly to drinking water pollution. Water quality degradation has far-reaching consequences, including harming health, the environment, and infrastructure. Waterborne infections kill more than 1.5 million people (about the population of West Virginia) each year, according to the United Nations (UN), much more than accidents, crimes, and terrorism. As a result, it is extremely crucial to forecast the quality of water. Earlier, water quality was checked manually. Yet, systems taught with machine-learning methods such as Linear Regression and SVM (Support Vector Machines) classifiers are later employed independently. The present implementation predicts the quality of water utilizing different Supervised Machine Learning methods such as Linear Regression, K-Nearest Neighbour, Decision trees, XGBoost, and Random Forest trees. Finally, this study seeks the algorithm that provides the highest accuracy while still maintaining water quality. This study compares multiple Machine Learning algorithms in tracking the water purity through KNN, Decision tree, Random Forest trees, SVM, and Gradient Boosting Classifier. The Water Quality Index dataset from Kaggle was used to train this model.

Description:

Water Quality prediction; Supervised Machine Learning; KNN; SVM; Decision Tree; XGBoost; Random Forest trees