Comparative analysis of the K-Medoids clustering algorithm and self organizing maps for Indonesia labor market indicator in 2017 – 2019.

Clustering analysis is an analysis that aims to place a set of objects in two or more groups based on the similarity of their characteristic objects. There are several clustering algorithms, namely the K-Medoids Algorithm and Self Organizing Maps (SOM). The case study in this research is related to employment, labor is one of the fields to advance the economy of a country in terms of quantity and quality as a target for developing markets and resources for the production and distribution of goods and services. Repairing labor problems in Indonesia using the labor market provides a way out to create conditions of synergy between sellers and employers, so it is necessary to implement good cooperation between the parties involved, namely labor sellers, labor buyers, and the government. Therefore, researchers are interested in conducting research related to employment in Indonesia in 2017-2019 using the K-Medoids Algorithm and Self Organizing Maps (SOM). Based on the results of the cluster using the K-Medoids Algorithm, the results obtained on the average data from 2017 to 2019, namely cluster one consists of 29 provinces and cluster two there are 5 provinces. While the results using the SOM algorithm are obtained on the average data from 2017 to 2019, namely cluster one consists of 28 provinces and cluster two there are 6 provinces. Furthermore, based on the results of the comparison of the best cluster method using internal validation (Connectivity Index, Dunn, and Silhouette) the best algorithm was obtained, namely the K-Medoids Algorithm. [ABSTRACT FROM AUTHOR]