<![CDATA[Population growth in West Nusa Tenggara Province continues to increase, making birth rate control through the Family Planning (KB) program an urgent priority. One key indicator of the program’s success is the pattern of contraceptive use across regions. This study aims to group districts in West Nusa Tenggara Province based on contraceptive usage patterns using two clustering methods: Gaussian Mixture Model (GMM) and Affinity Propagation (AP), and to compare the validity of the clustering results. The data used are secondary data on the number of contraceptive users in every district of West Nusa Tenggara Province for 2024, obtained from the SIGA BKKBN NTB website. Six types of contraceptives were included in the analysis: Injection, Pill, Implant, IUD, Vasectomy, and Tubal Ligation. Analysis was conducted using GMM, where the optimal number of clusters was determined using the Connectivity, Dunn, and Silhouette indices; and AP, which identifies the number of clusters automatically based on a similarity matrix. Model validation was performed using the average Silhouette Score, and results were visualized through Principal Component Analysis (PCA) and spatial mapping. Results show that GMM formed 2 clusters with a Silhouette Score of 0.412 — Cluster 1 comprising 32 districts and Cluster 2 comprising 85 districts. Meanwhile, AP produced 13 clusters with a Silhouette Score of 0.465, which is higher than that of GMM. This indicates that AP achieves better inter-cluster separation and is more optimal for grouping districts based on similarities in contraceptive use patterns. These findings can serve as a basis for BKKBN and local governments in formulating targeted, effective policy interventions for the Family Planning program.]]>
Copyrights © 2026
