Cluster Analysis Of Cities/Districts In West Kalimantan Based On Stunting Response Indicators Using The Calinski Harabasz Index

Tegar, Rama Priyatna (2025) Cluster Analysis Of Cities/Districts In West Kalimantan Based On Stunting Response Indicators Using The Calinski Harabasz Index. Skripsi thesis, Universitas Tanjungpura.

	Text (Cover-Bab I) Cover-Bab1_H1011211014.pdf - Published Version Download (994kB)
	Text (H1011211014_TEGAR RAMA PRIYATNA) H1011211014_TEGAR RAMA PRIYATNA.pdf - Accepted Version Restricted to Repository staff only Download (2MB)

Abstract

Stunting in West Kalimantan (Kalbar) province in 2023 reached 27% or 10% higher than the national target. The problem is that there is a lack of targeting in prioritizing priority areas to be addressed because the indicators are only based on population and the number of mothers giving birth. The purpose of the study was to explain the role of the modified K-Means and CHI methods in forming the best cluster and interpreting the condition of each cluster. The research object uses eight indicator variables obtained from the family information system website and BPS-Statistics Indonesia in 2023 which include the number of adolescents attending counseling, the number of informed consent, the number of complication cases, the number of aslocon expenditures, the number of alocon supplies, the population growth rate, the population density level, and life expectancy. The eight variables used were selected because they had been analyzed using PCA with the Kaiser approach and PVE, which represented 76.76% of the information divided into three main dimensions. The role of the modified method is the formation of clusters by analyzing the data units of each variable, the level of characteristics of each object through Euclidean distance, finding the value of the cluster center (centroid) then iterating until the results are consistent as a result of cluster formation while the role of CHI evaluates the number of clusters from the minimum number of one to the maximum of seven to obtain the best number of clusters. The best K = 5 clusters are obtained, with the status of cluster 1 being quite bad, with three objects, cluster 2 is very bad, with four objects, cluster 3 is good, with three objects, cluster 4 is very good, with three objects, and cluster 5 is good, with one object.

Actions (login required)

View Item