ANALISIS PERBANDINGAN SILHOUETTE COEFFICIENT DAN METODE ELBOW PADA PENGELOMPOKKAN PROVINSI DI INDONESIA BERDASARKAN INDIKATOR IPM DENGAN K-MEDOIDS

Abstract

Development is a process of change that is carried out consistently with the aim of improvement in all aspects of life with the prevailing values in society to achieve predetermined life goals. The quality of life of the community is measured by the Human Development Index (HDI) at the provincial level through three indicators, including: economic level, health, and education. K-medoids is a method used to group objects that contain outliers. In determining the optimal number of clusters using the Silhouette Coefficient method which has the advantage of determining the best number of clusters that can measure how close the relationship between objects and measure how far a cluster is separated from other clusters. The technique called the Elbow method is employed to ascertain the optimal quantity of clusters, which is done by examining the percentage outcomes derived from comparing the quantity of clusters that form an elbow at a certain point. Then cluster evaluation is carried out using the Davies Bouldin Index (DBI) as a comparison for cluster validation. The results of this study using the Silhouette Coefficient method produced the best cluster, namely 3 clusters with a Silhoutte Coefficient value of 0,3129 and a DBI value of 1,3184. Meanwhile, using the Elbow method produced the best cluster, namely 4 clusters with an SSE value of 54,5548 with a DBI value of 1,1754. So that the best cluster is 4 clusters with the Elbow method with the smallest DBI value.