PENERAPAN CLUSTERING TIME SERIES UNTUK MENGGEROMBOLKAN PROVINSI DI INDONESIA BERDASARKAN NILAI PRODUKSI PADI

Abstract

In fulfilling national food needs, rice plants are an important main commodity. The characteristics of various regions that differ from one another, greatly affect the production of rice produced. The purpose of this research is to cluster provinces in Indonesia by applying clustering time series analysis. The application of clustering time series analysis is based on rice production data for each province in Indonesia in the period 1968 – 2015. Clustering time series analysis with hierarchical and non-hierarchical methods results in the distribution of the same cluster members in the 3 optimal groups. On average, from the 3 clusters formed, the silhouette coefficient value is 0.64, so the clustering is categorized in Good Classification.

Keyword: cluster, rice, silhouette, time series.