Evaluasi Kinerja dan Implikasi Kebijakan: Analisis Klasifikasi Tingkat Keparahan Kasus HIV/AIDS di Jawa Barat Menggunakan Decision Tree (CART)

Abstract

This study aims to classify the severity of HIV/AIDS cases in West Java Province using the Decision Tree CART algorithm. The data used include the number of cases by region, gender, and age group. This study uses a quantitative method with the Decision Tree classification algorithm and 10-fold cross-validation validation technique to measure model performance. The results showed that the Decision Tree model had an accuracy of 96.15% with a Kappa value of 0.94, which indicates a very good level of accuracy and consistency in classifying the severity of cases. Variable analysis showed that the number of women and the productive age group (25-49 years) were the most significant variables in influencing the severity of cases, followed by the age groups 20-24 years and 50-70 years. The distribution of cases showed that the high category was more dominant in women, while men had a more even distribution in the low to high categories. These findings provide important implications for local governments to design more adaptive and evidence-based HIV/AIDS control policies. Recommendations from this study include increasing access to health services for women, education on preventive behavior for productive age groups, and focusing interventions on areas with high case severity.