Application of the K-Means Algorithm for Clustering Plantation Crop Production in Indonesia

Abstract

Plantation crop production is one of the main sectors in increasing people's income. Data on plantation products held by the Central Statistics Agency (BPS) is in the form of raw data, namely the production results of each province each year. This makes it quite difficult for the government to identify provinces that have the potential to produce crops. By clustering plantation crop production results, it will be easier for the government to identify provinces that have the potential to produce plantation crops. In this research there were 3 plantation crop production, namely coconut, coffee and cocoa. The data used is data for 2017 – 2021 which consists of 29 provinces. From the 3 plantation crop production, the data was collected using the K-Means Algorithm Data Mining technique. The results of this research are groupings which are divided into 2 clusters obtained from the Sum of Squared Error (SSE) calculation with a minimum value of 279261.63, namely low production and large production. Based on the results of the K-Means Algorithm calculations, it was found that coconut production had a small production cluster of 24 provinces, a large cluster of 5 provinces, for coffee a small production cluster of 23 provinces was obtained, a large production cluster was 6 provinces, and for cocoa a small production cluster was obtained of 23 provinces. , a cluster of 6 provinces.