Implementasi Algoritma Agglomerative Hierarchical Clustering untuk Mengklasifikasikan Penerima Bantuan Sosial RASTRA

Abstract

Poverty is a problem that is being faced by national development in Indonesia. Social assistance is one way for the government to overcome the problem. In order to help analyze gaps in the community the writer decides to do a cluster analysis that can be used to group the community into a particular class. Cluster analysis is performed using Agglomerative hierarchical clustering (AHC) with the method of linking single linkage, complete linkage, and average linkage. In this study used the poverty survey data of Kuala Sangkang village residents in 2018 and the data of recipients of the RASTRA assistance for Kuala Sangkang village in 2018. The results of the clustering process were 227 data, then tested using a confusion matrix. Based on the analysis and testing results it can be concluded that the process of clustering recipients of literary assistance can be formed in two clusters where for single linkage, cluster 1 consists of 222 and cluster 2 has 5 data members. Complete linkage produces cluster 1 members as much as 54 data and cluster 2 produces 173 data, average linkage produces 200 members for cluster 1 and 27 members for cluster 2. the average linkage method has better grouping results than the other two methods with accuracy reaching 54.19%, Error rate is 45.81%, and True positive / Recall rate is 92.11%.