Komparasi Algoritma Support Vector Machines dengan Algoritma Artificial Neural Network untuk Memprediksi Nilai Persetujuan Kredit Modal Kerja yang Diberikan Bank Umum

Abstract

Credit may be meant money provision orcollection that can be equavalent with that, basedon credit approval or loan agreement between bankand other party who oblige lender to pay off thedebt after specific terms period with interestexpenses. Commercial Bank is a bank that operateits business in conventional and or based onsyariah principle which is in operation provide inand out payment service. In this business operation,commercial bank provides loan/credit facility to thecustomer in Rupiah and foreign currency. Workingcapital credit is a credit used to finnance workingcapital purposes are depleted in one or several timethe production. For example: to buy raw material,salary, rent a building, purchase merchandise andso forth. Working capital credit approval providedby commercial bank need to predict because it hasincreased of credit provision provided bycommercial bank that can be used as measurementof economic growth and country stability or asmeasurement of economic growth indicator frommonetary sector by Bank of Indonesia. In thisresearch will conducted working capital creditvalue approval prediction will be provided bycommercial bank using support vector machinealgorithm that is compared with artificial neutralnetwork algorithm. From the result of testing onsupport vector machine algorithm using kernel dotproviding the accuracy result : 68,8% and RMSE :11928,594 and the result acquired using artificialneutral network algorithm providing the accuracyresult : 84,7% and RMSE : 5806,350. This resultshows that the best performance for workingcapital credit value approval provided bycommercial bank is artificial neutral networkalgorithm.

How Far Has Our MSMEs Credit Underwriting Assessment in Indonesian Commercial Banks Progressed?

“…The use of complementary predictors in the advanced analytics era could be influential in enriching generally accepted MSMEs credit assessment systems. As an example, Sopian et al, (2019) use an artificial neural network (ANN) and support vector machine (SVM) to predict approved working capital loan value in Indonesia and validate the model with 10-fold cross-validation. Another implementation was studied by Rizki et al (2017), as they used machine learning to detect firms that issued fraudulent financial statements (FFS) from 124 companies and 24 of them conducted financial statements fraud as signified with ANN and SVM from 9 statistically significant variables.…”

Analysis of SVM and Naive Bayes Algorithm in Classification of Nad Loans in Save and Loan Cooperatives

“…Research that applies SVM to predicting the value of working capital credit approvals states that SVM is able to produce an average accuracy of 69% in predicting the value of working capital loans approved by one of the commercial banks [8].Research that applies SVM to predicting credit risk states that SVM is able to produce an average accuracy of 80.95% in predicting 63 cases of loan applications at a bank in Palu City [9].…”