Penerapan Algoritma Decision Tree Pada Penentuan Penerima Program Keluarga Harapan di Desa Turirejo, Kedamean Gresik

Abstract

Based on a press release from the Ministry of Finance's Fiscal Policy Agency, the poverty rate in Indonesia was recorded to have decreased from 9.71% to 9.54% in March 2022. Even though there is pressure on commodity prices, the poverty rate shows a downward trend. Apart from economic growth, Indonesia has several programs launched by the government to alleviate poverty, such as the Family Hope Program. The Family Hope Program is a government social assistance program aimed at poor communities who are designated as beneficiaries of the Family Hope Program. However, in a summary of the audit results for the second semester of 2021, the Financial Audit Authority (BPK) found an error in the allocation of national social assistance (Bansos) which resulted in state losses of up to IDR 6.9 trillion. This is certainly a serious problem, for this reason a system is needed that can assist in the classification process of potential PKH assistance recipients. To carry out classification, you can use the Decision Trend algorithm with the ID3, C45 and Random forest models. To improve model accuracy results, a features selection method is needed. From the results of the classification process, the Random forest model with SMOTE and Features Selection has the highest accuracy of 91% and the system validity test is 91%. From these results, the system has the ability to assist in the classification process of PKH assistance recipients