PENERAPAN K-MEANS CLUSTERING UNTUK PEMETAAN PRODUKTIVITAS PADI DAN PREDIKSI PANEN DI KABUPATEN INDRAMAYU

Indonesia is a country where the majority of people consume rice, making the issue of rice harvest productivity a concern for many parties. Efforts to maintain food security, especially the availability of rice, are a shared responsibility, where efforts that can be made are to maintain and predict rice production results in areas that are National Rice Granaries. Indramayu is one of the national rice producing districts, which has 31 sub-districts with varying amounts of rice production. The large difference in rice production between sub-districts is an important problem that affects rice harvests in Indramayu, so improvements need to be made, including clustering, namely grouping sub-districts based on harvest potential using K-Means clustering by looking at historical harvest data for the past 5 years between 2019–2023. Then compared using the orange data mining application. K-Means clustering either manually or using orange3 produces three clusters with the same sub-district cluster results. This data is also used to predict future harvest results using the random forest algorithm in Orange3. Analysis of K-Means clustering results and prediction results shows that sub-districts in the lower cluster have the potential for a greater percentage reduction in harvest than clusters with a higher harvest category.