Multivariable Panel Data Cluster Analysis using Ward Method Gross Enrollment Ratio (GER) Data in West Java in the Year 2015-2018

Abstract

The aim of this study is to determine cluster analysis for panel data with multivariable data structures. Choosing a Ward method Choosing a method in a cluster analysis hierarchical technique. Ward method is a method based on Sum Square Error (SSE) with a measure of homogeneity between two objects based on the minimum number of error squares. The measure of similarity used is the Euclidean distance squared. The Ward method is used to add variation between objects in one cluster and maximize variation with objects in another cluster. The steps of the analysis are described in the discussion of this study. The method of implementation uses education gross enrollment rate (GER) data in West Java Province in 2015-2018. The results of the study indicate that the grouping of education GER data in West Java in 2015-2018 using the Ward method produces four clusters. The first cluster consists of five regions, GER for Elementary school, junior and senior high school in the cluster are below the average APK in West Java. The second cluster consists of two regions, in contrast to the first cluster GER for elementary schools in this cluster according to the average GER in West Java but for junior and senior high school GER below the average GER in West Java. The third cluster consists of seven regions, the GER for elementary, junior and senior high schools in this cluster is above the average GER in West Java while the fourth cluster consists of five regions, the GER for elementary and junior high schools in this area is above the average GER in West Java.