<![CDATA[Education is crucial for improving human resource quality. This study compares the K-Means and K-Medoids algorithms to analyze the distribution of education levels in Bekasi City. The dataset used was derived from 56 urban villages during 2021–2023, which were averaged to provide a more stable picture. The clustering process was performed using the RapidMiner application, with the results evaluated based on the Davies Bouldin Index (DBI) to measure the validity of the data groups. The evaluation results showed that the K-Means algorithm produced better performance than K-Medoids, indicated by the lowest DBI value of 0.558 at k = 4, while the K-Medoids algorithm produced the best DBI value of 0.803 at k = 3. The lower DBI value in K-Means indicates better cluster quality, thus K-Means was chosen as the best method for mapping education levels in Bekasi City. In addition, the K-Means results are also consistent with the Elbow method which shows the optimal number of clusters at k = 4, while K-Medoids shows a discrepancy between the results of the Elbow Method (k = 5) and DBI (k = 3). The K-Means clustering results form four clusters with different educational level characteristics. Cluster 0 is dominated by secondary education so that it requires scholarship programs and vocational training. Cluster 1 has a higher level of education so that it needs to focus on developing further education infrastructure and creating quality jobs. Cluster 2 has a low to middle level of education so that it requires increasing access to education and extension programs. Meanwhile, Cluster 3 shows a relatively even distribution of secondary to higher education so that it requires improving the quality of education and professional training to support human resource development in Bekasi City.]]>
Copyrights © 2026
