Answers are grounded in this paper and cite their sources.
Observational analysis classifies social assistance recipients using exhaustive chaid, indicating potential targeting improvements.
The poverty rate in West Java in 2024 reached 7.78% or around 3.8 million people. The distribution of social assistance through the Family Welfare Card (KKS) program still suffers from inaccuracy in targeting, with a potential loss to the state of Rp523 billion per month according to the BPK. This study uses the Exhaustive CHAID method to accurately classify KKS recipient households. The data used comes from the 2024 SUSENAS survey with 13,861 respondents, divided into 11,045 training data and 2,816 testing data, involving 12 independent variables. The resulting decision tree has a depth of three levels and 10 terminal nodes. The five most significant variables are the highest level of education of the household head (KRT), the number of health insurance policies, the type of floor material, the number of household members covered, and the status of home ownership. Households with low education levels, health insurance, and semi-permanent floors tend to receive KKS. Conversely, households with high levels of education, clear ownership status, and relatively few members were more likely to be classified as not receiving KKS. Model performance evaluation shows an accuracy of 64.72% on the training data and 66.23% on the testing data. The precision, recall, and F1-score values on the testing data are 57.3%, 57.76%, and 57.53%, respectively. The Matthews Correlation Coefficient (MCC) value of 0.2950 indicates a positive correlation, although it is still low, possibly due to an unbalanced class distribution. These results suggest that Exhaustive CHAID is sufficiently effective in classifying KKS recipients and can be used to improve the accuracy of social assistance distribution. Abstrak. Tingkat kemiskinan di Jawa Barat tahun 2024 mencapai 7,78% atau sekitar 3,8 juta jiwa. Penyaluran bantuan sosial melalui program Kartu Keluarga Sejahtera (KKS) masih mengalami ketidaktepatan sasaran, dengan potensi kerugian negara sebesar Rp523 miliar per bulan menurut BPK. Penelitian ini menggunakan metode Exhaustive CHAID untuk mengklasifikasikan rumah tangga penerima KKS secara tepat. Data yang digunakan berasal dari SUSENAS 2024 dengan 13.861 responden, dibagi menjadi 11.045 data training dan 2.816 data testing, serta melibatkan 12 variabel independen. Pohon keputusan yang dihasilkan memiliki kedalaman tiga tingkat dan 10 terminal node. Lima variabel paling signifikan yaitu pendidikan terakhir kepala rumah tangga (KRT), jumlah jaminan kesehatan, bahan lantai rumah, jumlah anggota rumah tangga ditanggung, dan status kepemilikan rumah. Rumah tangga dengan pendidikan dasar, memiliki jaminan kesehatan, dan lantai semi permanen cenderung menerima KKS. Sebaliknya, rumah tangga dengan pendidikan terakhir KRT yang tinggi, status kepemilikan rumah tangga nya jelas, serta jumlah anggota rumah tangga yang relatif sedikit, lebih dominan diklasifikasikan sebagai tidak menerima KKS Evaluasi performa model menunjukkan akurasi sebesar 64,72% pada data training dan 66,23% pada data testing. Nilai precision, recall, dan F1-score pada data testing masing-masing sebesar 57,3%, 57,76%, dan 57,53%. Nilai Matthews Correlation Coefficient (MCC) sebesar 0,2950 menunjukkan adanya korelasi positif meskipun masih rendah, kemungkinan karena distribusi kelas yang tidak seimbang. Hasil ini menunjukkan bahwa Exhaustive CHAID cukup efektif dalam mengklasifikasikan penerima KKS dan dapat digunakan untuk meningkatkan ketepatan penyaluran bantuan sosial.
No takes yet. Share an insight, caveat, or question.
Maharani et al. (2025) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: