Analysis of Data Mining in the Group of Water Pollution Areas using the K-Means Method in Indonesia

Abstract

Water is an essential requirement in human life. However, pollution causes the water quality to become poor, making it unsuitable for use. Pollution comes from rubbish and waste dumped into rivers, lakes and other water areas. This study aims to carry out a model for mapping areas contaminated with water pollution using artificial intelligence techniques. The data sample comes from the Indonesian Central Statistics Agency (abbreviated as BPS) which consists of 34 records. The data used are provinces in Indonesia that are polluted by water pollution in rural areas. The intelligence technique used is data mining using the k-means method. The variable used is the number of polluted villages by province. The mapping label used is the high cluster (K1) for water pollution and the low cluster (K2) for water pollution. Analysis using Rapid Miner software. The results showed that 4 provinces were included in the high cluster (K1) category, namely North Sumatra, West Java, Central Java, East Java. Testing of cluster results was carried out using Davies Bouldin (k = 2) with a value of 0.328, which means that the cluster results created were optimal. The results of the analysis are expected to be input for the government in focusing on areas polluted with water pollution.