The Use of K-Means Algorithm Clustering in Grouping Life Expectancy (Case Study: Provinces in Indonesia)

Abstract

Life expectancy is defined as information that illustrates the age of the death of a population. Life expectancy is a general picture of the state of a region. If the infant mortality rate is high, then the life expectancy in the area is low. And vice versa, if the infant mortality rate is low, the life expectancy in the region is high. Life expectancy is also a benchmark for government actions in improving the welfare of society and the human development index. For this reason, it is necessary to group life expectancy data to make it easier to determine the provinces with high, middle, and low life expectancy. The results of cluster testing using the silhouette score method showed that two subjects had a low silhouette score level, which caused the cluster value to be less than optimal, namely East Java  & Gorontalo. The clustering results found that the cluster was divided into 3, namely cluster 1, with a high level of life expectancy consisting of 10 provinces, namely East Java, Riau, North Sulawesi, Bali, North Kalimantan, DKI Jakarta, West Java, Central Java, East Kalimantan and Special Region of Yogyakarta. Cluster 2 has a level of middle-life expectancy consisting of 18 provinces, namely Gorontalo, North Maluku, Central Sulawesi, South Kalimantan, North Sumatra, Bengkulu, West Sumatra, Central Kalimantan, Aceh, South Sumatra, Banten, Kep. Riau, South Sulawesi, Kep. Bangka Belitung, Lampung, West Kalimantan, Southeast Sulawesi and Jambi. Cluster 3, with a low level of life expectancy, consists of 6 provinces, namely West Sulawesi, Papua, Maluku, West Papua, West Nusa Tenggara, and East Nusa Tenggara.