Hierarchical and K-means methods for analyzing the Indonesian welfare indicator in the year 2020

Abstract

The main goal in the development of a country is to improve the welfare of the people. One of the causes of people's welfare problems is that the development carried out by the government is not carried out properly. The solution that can overcome this problem is by identifying what indicators that can affect the level of people's welfare. In this research, we investigate the effect of indicators of Life Expectancy, Average Length of Schooling, Percentage of Poor Population, and Open Unemployment Rate on the level of welfare in Indonesia using cluster analysis. Cluster analysis is a technique of grouping objects based on similar characteristics that they have. The methods proposed in this research are hierarchical cluster method which is Average Linkage and non-hierarchical method which is K-Means. From the results of the analysis, it was found that 2 clusters were formed for each method. There were 4 provinces included in cluster 1 and 30 provinces included in cluster 2 using the Average Linkage method. As for the K-Means method, 12 provinces are included in cluster 1 and 22 provinces are included in cluster 2. Based on the average of each indicator of people's welfare, we conclude that cluster 1 is a low cluster and cluster 2 is a high cluster for the Average Linkage method and K-Means method.