Prediction of K-Means Multivariable Rice Production Center Pattern

Abstract

West Sumatra is one of the provinces that have the potential to be used as a rice production center in Indonesia. However, the determination of local production centers is still justified based on conventional methods. This study aims to obtain historical data on rice production center variables, process data using the K-Means algorithm (RStudio), and justify rice production center areas in West Sumatra in a given year based on the acquisition of recommendation values (centroids). The results showed that in 2010-2020, the best number of clusters (elbow graph) was K=3. The highest recommendation value of the final centroid lies in centroid 3 so 8 recommendations for districts that are the center of rice production are obtained, including Padang Pariaman Regency, Pesisir Selatan Regency, Tanah Datar Regency, Agam Regency, Solok Regency, Fifty Cities Regency, Pasaman Regency, and South Solok Regency. Furthermore, of the eight districts included in the Governor’s Decree no. 525-757-2021 concerning the determination of food areas. Based on the 2019 irrigation network performance index as a decision-supporting variable, the highest index value reached the Batang Surantih irrigation area, Pesisir Selatan Regency, with an index value of 15 (100%).