Application of K-Medoids Algorithm in Provincial Grouping in Indonesia Based On Case of Environmental Pollution

Abstract

Cluster analysis is a method for grouping objects that have the same characteristics. One of the methods in cluster analysis used to group data is the K-Medoids method. In this study the K-Medoids method was applied to classify provinces in Indonesia based on environmental pollution. The variables used are: the number of sub-districts/villages that experience water pollution from factory waste, the number of sub-districts/villages that experience water pollution from household waste, the number of sub-districts/villages that experience soil pollution from factory waste, the number of sub-districts/villages that experience soil pollution from household waste, the number of sub-districts/villages that experience air pollution from factory waste and the number of sub-districts/villages that experience air pollution from household waste. Based on the Davies Bouldin Index, the 2 best clusters were obtained where the first cluster consisted of 31 provinces which had low environmental pollution and the second cluster consisted of 3 provinces which had high environmental pollution.