K-Means Clustering Method for District/City Grouping in Effort to Control Inflation Rates in Java and Sumatera

Abstract

Indonesia is developing Special Economic Zones (SEZs) in order to develop the economy after COVID-19. The islands of Java and Sumatra is the highest number of SEZs and have with economic growth tend to dominate in Indonesia. Basically, the level of the economy is expected to tend to be stable and or increase. However, inflation can affect the level of economic growth. The mapping of regions based on inflation rate carried out in this research can be one of the efforts to control inflation. One of the statistical methods for grouping regions is k-means clustering. This research uses secondary data from the BPS. The research result showed that three cluster groups were formed based on expenditure indicators, which are 25 regions in the first cluster, 21 regions in the second cluster, and 4 regions in the third cluster. Based on the three clusters formed, the characteristics of the indicators of food, beverages, and tobacco; housing, electricity, and home fuel; health; recreation, sport, and culture; and the provision of food and beverages (restaurants) have a significant difference in each cluster. So, in formulating policies related to inflation control, the government is advised to make different policies for each cluster related to these indicators. The government also needs to make policies to maintain the level of public consumption on expenditure indicators that tend to be high in each cluster, so it is hoped that inflation rate for these expenditure indicators is stable and can be controlled more effectively and efficiently.