Mencegah Kredit Macet Dengan Analisa Kelayakan Pembiayaan Dengan Metode C4.5 Dan Naïve Bayes (Studi Kasus : Koperasi BMT UGT Sidogiri Cabang Cikarang)

Abstract

The progress of the growth of MSMEs (Micro, Small and Medium Enterprises) in Indonesia from time to time which is increasingly rapid, resulting in an increase in the need for capital to develop their business. This is evidenced by the increasing number of credit or financing withdrawals from savings and loan cooperatives and BPRs (Rural Banks). The problem faced by savings and loan cooperatives, BPRs, or other financial institutions at this time in providing credit is the risk of late payments, repayments and even failure of credit payments. This problem occurs due to credit misuse and weak supervision both in the process of providing credit and in the implementation stage. The right solution to solve existing problems is by using data mining algorithms. The concept of data mining will make it easier to solve problems that are not optimal in cooperatives, the classification method is able to find models that differentiate concepts or data classes with the aim of making it easier to predict creditworthiness. The Naive Bayes algorithm and the C4.5 algorithm are considered to be able to predict future opportunities based on previous experiences. The author conducted research on the BMT UGT Sidogiri Cooperative with the title "Preventing bad credit by analyzing the feasibility of financing with the Naive Bayes and C4.5 methods". In this study the authors used 9 attributes as an assessment, namely: name, residence status, financing contract, income, ceiling, term of repayment, number of dependents, collateral. Testing is done using 520 data and 104 randomly selected testing data. From the results of tests carried out using Rapid Miner tools, it can be concluded that the accuracy level of the C4.5 algorithm is more accurate at 81.35%, while the Naive Bayes algorithm is 78.85%.

Keywords : Credits, Classification, Accuracy, Naive Bayes, C4.5.