CLUSTERING ANALYSIS WITH AVERAGE LINKAGE METHOD FOR GROUPING PROVINCE IN INDONESIA BASED ON WELFARE INDICATORS

The high level of social inequality in Indonesia is a problem that must be resolved immediately. High social inequality will result in an increase in social tension which also impacts on the high level of conflict and crime in society. The problem of social inequality can be solved by accelerating the welfare distribution program by the government. The provision of this program must be fair and adapted to the conditions needed by each region. This is because each region has different causes of welfare problems. Therefore, in providing the program, the government must have a priority scale on welfare issues in an area that can be done using a mathematical method in the field of statistics, namely cluster analysis. This study aims to obtain, analyze and interpret the results of grouping provinces in Indonesia based on indicators of people's welfare. As many as 34 provinces in Indonesia as objects will be grouped based on 20 variables related to people's welfare. The grouping is done using the Hierarchy Method, the agglomeration grouping procedure with the Average Linkage technique and the size of the Euclidean Distance. From the clustering algorithm, it was found that from 34 provinces in Indonesia grouped into 5 clusters namely Cluster 1 consisting of 24 members namely Aceh Province; North Sumatra; West Sumatra; Riau; Jambi; South Sumatra; Bengkulu; Lampung; Head of Pacific Islands; Riau Islands; West Java; Central Java; East Java; Banten; West Nusa Tenggara; Central Kalimantan; South Borneo; East Kalimantan; North Kalimantan; North Sulawesi; Central Sulawesi; South Sulawesi; Southeast Sulawesi; and Gorontalo. Cluster 2 consists of 1 member, DKI Jakarta Province. Cluster 3 consists of 2 members including DI Yogyakarta and Bali Provinces. Cluster 4 consists of 6 members including East Nusa Tenggara Province; West Kalimantan; West Sulawesi; Maluku; North Maluku; and West Papua. Cluster 5 consists of 1 member, Papua Province. Based on the comparison of the average value of each cluster, the five clusters are sorted based on their level of welfare, namely: Cluster 3 as a very good cluster, Cluster 2 as a better cluster, Cluster 1 as a good cluster, Cluster 4 as a pretty good cluster and cluster 5 as a less good claser.   Keywords: Cluster Analysis, Average Linkage, People's Welfare &nbsp