Data mining tools | rapidminer: K-means method on clustering of rice crops by province as efforts to stabilize food crops in Indonesia

Indonesia is an agriculture-based country. Agriculture is a sector that becomes the backbone of Indonesia's economic development and improvement. Food security is one of the most important of farms. There are several types of food crops that become important commodities for the nation of Indonesia namely rice, corn, peanut, green beans, cassava and sweet potatoes. This research data is sourced from BPS-Statistic (https://www.bps.go.id/). This research raised the topic of rice crops clustering by province (1993-2015) using data mining with K-means method. The method used with the help of rapidminer software. The sample data used are 34 provinces in Indonesia with 3 parameters, namely: 1). Lack of Harvest Area (hectares), 2). Productivity (quintal / hectare) and 3). Production (ton). Cluster results using 3 clusters: (C1) high production cluster, (C2) normal production cluster and (C3) low production cluster. Based on the research results obtained (C1) high production cluster = 3 provinces, (C2) normal production cluster = 23 provinces and (C3) low production cluster = 8 provinces. This study also uses "% performance" to see the accuracy of the algorithm used with the research topic. From the result of accuracy using parameter average within centroid distance and Davies Bouldin obtained Davies-Bouldin index for rice plant is -0.392. Based on these performance results can be summed up as the best algorithm based on criteria. The lowest cluster clustering (C3): Aceh, North Sumatera, West Sumatera, South Sumatera, Lampung, West Nusa Tenggara, South Kalimantan, and South Sulawesi are input inputs to the government, to provide socialization to the province to increase rice production, is one of the commodities of Indonesian people, especially rice.