<![CDATA[This research is motivated by the high prevalence of stunting in Indonesia, reflecting nutritional imbalances in early childhood. To address this issue, an information technology approach is employed to identify at-risk infant groups. The analyzed data consists of anthropometric information, including height, weight, and age of infants, collected from the Peusangan Health Center. The applied method is the Gaussian Mixture Model (GMM) with the Expectation-Maximization algorithm to cluster the data into two groups: "Potential Stunting" and "Not Stunting." The research results indicate that several Posyandu and villages have notably high potential stunting rates, such as Posyandu Bungong Seulanga (141 infants) and Pante Gajah village (116 infants), with a higher prevalence among male infants (34.67%) and those aged 52–60 months (24.18%). Model evaluation using a confusion matrix on 1,465 data points showed a True Positive of 958 (65.36%), False Negative of 4 (0.27%), False Positive of 503 (34.33%), and True Negative of 0 (0%), with an accuracy of 65.36% and an error rate of 34.64%. However, a previous accuracy test on 1,665 data points only achieved 34.55%, indicating unsatisfactory individual prediction performance. In conclusion, Mixture Modelling is effective for clustering and identifying at-risk groups but lacks accuracy in individual predictions, with a bias toward the "Potential Stunting" class that requires improvement in future research.]]>
Copyrights © 2025
