Cluster Analysis of Inclusive Economic Development Using K-Means Algorithm

Abstract

This study aims to cluster 38 Districts/Cities in East Java based on the 10 forming indicators of inclusive economic development and to determine the inclusive economic growth of Districts/Cities above or below the total average. 10 indicators used in this study are GRDP per capita, GRDP by business field, Labor force participation rate, Unemployment rate, Gini ratio, Expenditure per capita, the number of poverty, Life expectancy, expectation years of schooling, and mean years of schooling. There are 3 scenarios in this study, namely 2 clusters, 3 clusters, and 4 clusters. The method of clustering in this study is using the K-means algorithm. This study uses the silhouette coefficient to evaluate the best cluster of 3 scenarios. The best k-means algorithm in this study is using 2 clusters with a silhouette coefficient of 0.87. There are 29 Districts/Cities included in cluster 1 with inclusive economic development below the total average and 9 Districts/Cities included in cluster 2 with inclusive economic development above the total average. The members of cluster 1 are mostly district areas and located in coastal or border areas and the members of cluster 2 are mostly urban or industrial areas.