Divorce Pattern Clustering Using The K-Prototype

Abstract

Divorce is a complex social problem that continues to increase every year, including within the jurisdiction of the Banyuwangi Religious Court. Various factors such as early-age marriage, economic problems, prolonged conflict, and domestic violence are the main triggers of divorce. This study aims to classify the divorce patterns of the Banyuwangi community using the K-Prototype clustering method, which is capable of handling mixed numerical and categorical data. The data used are secondary data from the Case Tracking Information System (SIPP) of the Banyuwangi Religious Court in 2025, totaling 5,570 cases. The variables analyzed include the age of the plaintiff and defendant, number of children, length of marriage, type of case, occupation, education, and divorce factors. The clustering process was carried out through preprocessing stages, determining the gamma parameter, data encoding, and implementing the K-Prototype algorithm using the Python programming language and the Streamlit framework. The results of the study show the formation of two main clusters. Cluster 0 (59.7%) is dominated by young couples with an average plaintiff age of 29.8 years, a short marriage duration of 6.3 years, and an average of 0.7 children. Cluster 1 (40.3%) is dominated by mature couples with an average plaintiff age of 45.5 years, a long marriage duration of 17.4 years, and an average of 1.2 children. PCA visualization and scatter plots further emphasize the distinction between the two clusters based on age. The developed SKPP application successfully facilitates users in uploading data, preprocessing, clustering, and exporting results. Therefore, the K-Prototype method is effective for analyzing divorce patterns and can serve as a supporting tool for the Religious Court in formulating more targeted prevention policies.