Optimalisasi Data Tidak Seimbang Pada Data Nasabah Koperasi dalam Pemberian Pinjaman Menggunakan Random Oversampling

Abstract

Cooperatives have developed from time to time, in providing services, credit cooperatives certainly have certain requirements as prospective customers to receive loans. Cooperatives need to check whether interested parties will receive loans. Loans to customers are the main source of income for cooperatives. In data mining, there are several classification algorithms that can be used for credit analysis, including the Random Forest and the C4.5 Algorithm. Data on prospective customers received from cooperatives as a condition for applying for credit is processed using Random Forest data mining and C4.5 Algorithm to support credit analysis in order to obtain accurate information on whether the prospect who applies for credit is feasible or not, this study was conducted to classify loans to prospective customers. cooperative customers using the Random Forest method and the C4.5 Algorithm which is optimized by Random Oversampling because the dataset is in an unbalanced condition. In testing the C4.5 Algorithm which is optimized with Random Oversampling, it gets an accuracy of 78.03%, where the accuracy increases by 7.89% from the previous 70.14%. Meanwhile, Random Forest with Random Oversampling has an accuracy value of 87.12%, an increase of 23.69% from the previous Random Forest test of 63.43