<![CDATA[Abstract−The assessment of outstanding urban villages at the city level in Jambi faces challenges in defining categories for coaching groups and selecting the appropriate algorithm to handle heterogeneous data. To address this issue, clustering techniques using K-Means and Fuzzy C-Means algorithms were applied to segment coaching needs based on homogeneous value proximity. This study analyzed 204 sub-district assessment data using Python, producing clustering results with nearly equivalent cluster separation quality. The K-Means algorithm achieved a silhouette score of 0.3854, slightly higher than Fuzzy C-Means at 0.3831. Both algorithms formed consistent cluster patterns with average values of 82, 72, and 61, classified into three clusters: (1) Cluster 0 receives awards and career development promotions, (2) Cluster 1 focuses on management training and bureaucratic reform, and (3) Cluster 2 requires coaching clinics and technical guidance. The findings indicate that K-Means is more advantageous due to its simplicity, effectiveness in handling linear datasets, and clear data distribution. This clustering approach supports the Jambi City Government, particularly the Regional Secretariat Governance Section, in designing data-driven coaching strategies to enhance the quality of sub-district development.]]>
Copyrights © 2025
