A Study of KNN Classifier to Predict Water Pollution Index

Abstract

Pattern recognition in data mining is the process of recognizing patterns by using machine learning algorithm. Data are classified based on the knowledge and represented after extracting patterns. Direct sources of drinking water are rivers, lakes and dams. Consuming safe drinking water is a fundamental need as well as human right. Prior to its use for drinking, water quality should be examined to check whether it is free from contamination. KNN, a pattern recognition classifier, is used for regression and classification problem. KNN uses a dataset and classifies data points based on similarity measures. It helps in quality predictions in most of the applications. Since drinking water may consist of various parameters in varying proportion, investigating the proportion is the need. This paper reviews the current status of drinking water and basics of KNN classifier. Further, it also studies the use of k-nearest neighbor classifier to predict and measure the accuracy of the proportion of parameters available in terms of the quality index of drinking water.