<![CDATA[Disparities in toddler immunization coverage across different regions of Indonesia indicate the need for an analytical approach capable of capturing regional characteristics in greater depth. This study aims to cluster toddler immunization coverage using the K-Means algorithm based on the 2025 Indonesian People’s Welfare Data published by Statistics Indonesia. The variables analyzed include immunization history, types of immunization, place of immunization, and immunization providers. Data processing was conducted using Python through the Google Colab platform. The determination of the optimal number of clusters using the Elbow Method resulted in three clusters, with a Silhouette Score of 0.5086, indicating a moderately good clustering quality. The results show that the cluster labeled Low surprisingly exhibits the highest immunization coverage (95.68%), suggesting that the cluster labels do not represent coverage levels in a linear manner but instead reflect differences in operational characteristics and the distribution of immunization services across regions. Meanwhile, the Medium cluster shows the lowest coverage (63.82%), and the High cluster falls at an intermediate level (92.23%). Further analysis indicates that the type of immunization and immunization history are the most influential variables in cluster formation. With clustering quality categorized as moderately good, the K-Means method is considered capable of adequately identifying immunization coverage patterns for region-based policy analysis. These findings demonstrate that a clustering approach can reveal immunization coverage patterns that are not captured through conventional statistical analysis and can serve as a basis for more targeted and data-driven immunization policy formulation.]]>
Copyrights © 2026
