Water Potability Analysis and Prediction

Abstract

Water is one of the largest resources on earth. People need water to sustain life, including drinking water. It is important to know whether drinking water - human life resource - is enough for everyone now and in the future. However, water resources are not evenly distributed everywhere on the planet. While the water resource is rich in some countries and regions, it is not enough for some other regions. The analysis of different region’s water resources should be done individually. In this paper, the authors analyze the potability of water by using an Indian water potability dataset from Kaggle. More specifically, this paper talks about each factor of water that influences water potability through statistical methods - binomial distribution and the k-nearest neighbor algorithm. Also, the authors build a model that allows people to predict the potability of a water resource by the data of each factor of that resource. According to the research, the features of water are not related to each other. All the features should meet a specific standard in order to get potable water.

Performance analysis of samplers and calibrators with various classifiers for asymmetric hydrological data

“…A publicly available asymmetric hydrological dataset has been 6+-gathered and applied. The water potability dataset is available in a .csv file from Kaggle repository [40]. It comprises 3,277 observations and includes 10 parameters: potential hydrogen (pH), hardness, solids or total dissolved solids (TDS), chloramines, sulphate, conductivity, organic carbon, trihalomethanes (THMs), turbidity, and potability [4042].…”

“…The turbidity of water is determined by the quantity of suspended solid matter, and it reflects water's ability to transmit light. Water potability, typically expressed on a scale from 0 to 1, indicates its suitability for human consumption [40].…”