IMPLEMENTATION OF THE K-NEAREST NEIGHBOR METHOD IN KNOWING WATER QUALITY

Abstract

The function of classification is the process carried out in predicting a data that has a class that is still unknown, the pattern that is owned is also already regular in a classification method. K-NN is a group that has an instances-based learning system, in conducting group searches by performing the value of k objects into the test with the closest value to the value of other data. KNN uses the closest distance value to the tested dataset in carrying out the classification process. Drinking water is very important for health and is a very effective component for the health of the human body. Health is very influential on the country's economy, it is necessary to invest in water that is very beneficial for the community. This study conducted a search for accuracy of water quality with data as many as 3276 different bodies of water in order to know which water can be drunk and not drinkable. The results of the accuracy of the KNN classification model that can increase the level of accuracy better for the data used. So the research on water quality has an accuracy of 56.40% with 370 data on drinkable water. Researchers hope that accuracy can be improved again by combining the optimization of the classification model in future studies