Klasifikasi Penentuan Penerima Bantuan Sosial Tunai (BST) Menggunakan Algoritma C4.5 Di Desa Keramas, Gianyar, Bali

The current Covid-19 pandemic has led to an increase in the poverty rate in Indonesia. To overcome the decline in income, the Ministry of Social Affairs provides Cash Social Assistance (BST) to 9 million KPM in Indonesia. One of the villages that received cash social assistance is Keramas Village. The problem is that the BST recipient is not precise. The purpose of this study is to classify the eligibility of BST recipients using the C4.5 algorithm and the C4.5 algorithm is expected to be able to provide recommendations for decision making in receiving other assistance. The results of this study are to predict whether the community is eligible or not eligible to receive BST. Researchers tested using 5 attributes, including having received other assistance, employment, education, having a poor card, and number of dependents. The test uses 2,074 data which is divided into 2 types of data, namely 80% training data and 20% testing data. This test produces an accuracy of 97.83% and compares with the K-Nearest Neighbor algorithm which produces an accuracy of 92.29% and Naïve Bayes of 91.81%.The current Covid-19 pandemic has led to an increase in the poverty rate in Indonesia. To overcome the decline in income, the Ministry of Social Affairs provides Cash Social Assistance (BST) to 9 million KPM in Indonesia. One of the villages that received cash social assistance is Keramas Village. The problem is that the BST recipient is not precise. The purpose of this study is to classify the eligibility of BST recipients using the C4.5 algorithm and the C4.5 algorithm is expected to be able to provide recommendations for decision making in receiving other assistance. The results of this study are to predict whether the community is eligible or not eligible to receive BST. Researchers tested using 5 attributes, including having received other assistance, employment, education, having a poor card, and number of dependents. The test uses 2,074 data which is divided into 2 types of data, namely 80% training data and 20% testing data. This test produces an accuracy of 97.83% and compares with the K-Nearest Neighbor algorithm which produces an accuracy of 92.29% and Naïve Bayes of 91.81%