Water Quality And Potability Using Machine Learning

Abstract

 Access to clean water to drink is crucial for good health, a fundamental human right, and an essential aspect of any health 
protection strategy. On a national, regional, and local level, this is significant as a development and healthcare issue. Access to 
safe drinking water is a global challenge. Traditional methods for water quality assessment can be time-consuming or require 
specialized equipment. This study explores the application of machine learning (ML) for portable water quality prediction. As a 
result, monitoring water quality now heavily relies on modelling and forecasting water quality. In this work, Machine learning 
algorithms are developed namely Decision tree, Random Forest, and Naïve Bayes algorithms for the dataset to predict the safety 
for human consumption. The study begins by compiling a comprehensive dataset comprising diverse water quality indicators 
such as pH levels, dissolved oxygen, turbidity, heavy metal concentrations, and microbial contaminants. Historical data from 
different sources, including public databases and local monitoring stations, are collected and preprocessed to ensure consistency 
and reliability. The used dataset has significant parameters and the developed model will be evaluated based on some statistical 
parameters. The goal of this algorithm is to create a model that predicts the value of a target variable 
(Decision tree, Random Forest, and Naïve Bayes score) and whether the water is potable or not