Algoritma K-Medoids Untuk Prediksi Hasil Produksi Buah Kelapa Sawit Berdasarkan Curah Hujan

Abstract

Oil palm is a variety of plantation crops that have an important role in the agricultural sector. Even so, there are several problems that are still unknown, one of which is the effect of rainfall on oil palm fruit yields. To understand the relationship between rainfall and oil palm yields, the clustering method uses the K-Medoids algorithm. The clustering method with the K-Medoids algorithm is used to classify oil palm yield data based on the rainfall that occurs. The purpose of using this algorithm is to determine the highest level of yield or production at PT. Sarana Titian Permata 2. In this study, the clustering results showed that there were four clusters produced. Based on performance testing, the best cluster chosen is 4 clusters. The selection of this cluster is based on the lowest Davies Bouldin Index (DBI) value obtained, which is -0.773 and the cluster results are obtained from data that has been grouped into four clusters. cluster 0 consists of 39 data, cluster 1 consists of 27 data, cluster 2 consists of 17 and cluster 3 consists of 13 data. By selecting this cluster, it is possible to identify oil palm yield groups that have better performance in relation to rainfall. This research provides a better understanding of the relationship between rainfall and oil palm yields in 2022 at PT. Sarana Titian Permata 2. By knowing the best clusters, efforts can be made to increase the productivity and efficiency of palm oil production based on existing rainfall conditions.