ANALISIS PENGELOMPOKAN DAERAH MENGGUNAKAN METODE NON-HIERARCHICAL PARTITIONING K- MEDOIDS DARI HASIL KOMODITAS PERTANIAN TANAMAN PANGAN (Studi Kasus Kabupaten/Kota Se-Jawa Tengah Tahun 2009 – 2013)

Abstract

Non-Hierarchical K-Medoids Partitioning is a clstering method for classifying objects based on the characteristics possessed by the object, wherein the object k randomly selected to be medoids is the center of the cluster. After medoids selected then other objects that have similarities with medoids made in one cluster. Medoids is the object which is considered to represent a cluster. Similarity between objects is calculated using euclidean distance. One application grouping method Non-Hierarchical K-Medoids Partitioning is to classify District in Central Java is based on the production of rice and pulses. Grouping Regency / City in Central Java using Non-Hierarchical Partitioning K-Medoids obtained information that rice production by Regency / City in Central Java can be grouped into seven clusters, but because of a case in 2010 and in 2011 the number of clusters that formed are two clusters, while the production of food crops by Regency / City in Central Java can be grouped into two clusters. Keywords: k-medoids, Non-Hierarchical, Euclidean distance, Similarities.