PENGELOMPOKAN PROVINSI DI INDONESIA BERDASARKAN KARAKTERISTIK KESEJAHTERAAN RAKYAT MENGGUNAKAN METODE K-MEANS CLUSTER

Abstract

Welfare have a relative explanation, dynamic, and quantitative. Quantitative formulation of welfare is never final because it will continue to evolve along with the development needs of human life. In 2011, the National Team for the Acceleration of Poverty Reduction (NTAPR) made priority sector that can serve as a benchmark the welfare in a region. From the priority sector will be made cluster or group which contains all 33 provinces based on the level of public welfare in the region uses data in 2012 were sourced from the Central Statistics Agency (CSA). The method that can be used to group the 33 provinces is K-Means Cluster method with number cluster as many as two, three, four, and five clusters. K-Means Cluster method is one of cluster analysis method who can partition the data into one or more clusters, so that the data with the same characteristics are grouped into the same cluster and data with different characteristics grouped into other clusters. To know the most optimal of the number of clusters we use Davies-Bouldin Index (DBI). We concluded that the optimal number of cluster is three with details the province in the first clusters have superiority in four sectors like net enrollment rate of primary school, net enrollment rate of junior high school, IMR (Infant Mortality Rate), and access to electricity. The province in the second clusters have superiority in one sector, that is open unemployment rate. The province in the third clusters have superiority in all sectors. Keywords : Welfare, NTAPR Priority Sector, K-Means Cluster Method, Davies-.Bouldin Index (DBI)