Determination of Granting Appropriateness Credit at “Daruzzakah Rensing” Cooperative Using the Support Vector Machine (SVM) Algorithm

Abstract

Credit is the main product of savings and loan cooperatives to increase profitability. The greater the credit issued, the greater the benefits obtained by cooperatives. Each cooperative will package credit products in such a way as to attract the attention of every customer. However, cooperatives can find problems in the process of lending, such as the “Daruzzakah Rensing” Cooperative located in “Desa Rensing, Kecamatan Sakra Barat, Lombok Timur-NTB-Indonesia”. The main products of the Cooperative “Daruzzakah Rensing” are savings and loans. In distributing credit, the cooperative always decides based on statistical data. This data is sometimes not useful if the supporting methods used to predict and classify the data are not appropriate. Therefore, this research requires a method that can classify and predict problematic and non-problematic customers. To answer this question, using the SVM (Support Vector Machine) algorithm to find out the level of accuracy in analyzing creditworthiness proposed by prospective debtors. The SVM algorithm is used to predict, classify, evaluate, and analyze credit. From the results of data processing carried out using the SVM algorithm (Support Vector Machine), it can be categorized as an excellent method, with an accuracy of 90.42% and AUC at 0.957. Accuracy of 90.42% means the SVM algorithm can provide decisions about feasible or not feasible in granting credit to customers who apply for loans.