Analyzing the Potability of Water using Machine Learning Algorithm

Human life cannot exist without water, as most individuals are taught at a young age. According to the World Health Organization (WHO), clean water is not readily available for one out of every six people on the planet, which is around 1 billion people. The potability of water is affected by many factors like pH, Hardness, and various other chemicals present in the water. Contamination in drinking water can cause severe diseases like Hepatitis A, Cholera, Dysentery, Typhoid, and Polio. The paper primarily focuses on the binary classification of water potability using various standard machine learning algorithms like K-Nearest Neighbors (KNN), XG-Boost, Decision Tree, Support Vector Machine (SVM), Logistic Regression, Gaussian Naïve Bayes, Random Forest, LightGBM, Bernoulli Naïve Bayes and Gradient Boost on selected features from the dataset by using correlation matrix and in terms of performance, Decision Tree outperformed all other machine learning algorithms by achieving an accuracy of 0.9358, Fl score of 0.9374 and AUC-ROC score of 0.9220 for classification of water potability.