Machine Learning Models for Water Quality Assessment and Prediction

Water is very crucial for every living being and utilizing of qualified water is much important. The solvent of water can have several components of polarity. Due to these compounds presented in water may be sometimes flavorless and unseen may leads to health problems for living beings. To reduce these risks much and should inspect the quality of a water. WQI (Water Quality Index) is hypothesis which gives discrete representation of standards of water at specific location. After applying WQI, if it fails to reach the standards, will take necessary actions before usage of non-standard water. In this study appraise of some machine learning strategies for assessing which includes Naïve Bayes, Random Forest, Decision Tree, AdaBoost, Logistic Regression, K-Nearest Neighbors (KNN), Support Vector Machine (SVM). The demographic data was gathered from publicly available kaggle repository with 21K samples. Performed comparison analysis among above algorithms, random forest out performances with 91 % accuracy and also provided which feature having the high importance useful for forecasting excellence of water.