Data mining approach for stunting clusters in Jumput Rejo

Abstract

The target of reducing the stunting prevalence rate by 14% in 2024 which has been set by the government needs to be of concern to be implemented by the local health office. The purpose of the research is to cluster toddler anthropometry data with data mining algorithm. Optimize K-Means (KM) algorithm with elbow method use to cluster toddler anthropometry data (sex, height, weight, age, and health care center). A set of 580 children's anthropometric measurements were analyzed and categorized based on their similarity. Cluster 1 comprises 150 members and exhibits a narrower range of age and height values compared to the other clusters. Cluster 2, with 124 members, displays a broader range of age and height values compared to both Cluster 1 and Cluster 3. Cluster 3, consisting of 150 members, demonstrates age and height values that are higher than Cluster 1 but lower than Cluster 2 and Cluster 4. Finally, Cluster 4, encompassing 156 members, exhibits age and height values that are higher than those in the other clusters that many children are stunted based on standard anthropometric table for assessing children's nutritional status. The cluster optimization yielded four distinct clusters, which will serve as the input for identifying clusters during the data grouping process using the KM algorithm.