Credit Collectibility Prediction of Debtor Candidate Using Dynamic K-Nearest Neighbor Algorithm and Distance and Attribute Weighted

Abstract

BPR Bank Jepara Artha is one of the banks that provide loan for activist of MSME (Micro, Small and Medium Enterprises). The activity of loaning in BPR Bank Jepara Artha has bad loan issue that often occured especially on loan MSME activist, therefore it needs an application to predict the loan Collectibility of debtor applicant to minimize the issue. This research applied one of Data Mining classification algorithms in the application that produces output that can serve as information sources or second opinion for the consideration in decision making to accept or reject the loan applicant. The algorithm that be used was Dynamic K-Nearest Neighbor and Distance and Attribute Weighted algorithm which is a dynamic selection of k, addition of attribute and distance weight on k-Nearest Neighbor algorithm. The attributes that be used to determine the prediction result are 5C (Character, Capacity, Capital, Collateral, Condition of Economic), monthly income, debt status elsewhere, number of dependents, age, type of commodity and business status. The results of Dynamic K-Nearest Neighbor and Distance and Attribute Weighted algorithm performance measurement use historical datas of 240 old customer, the order of importance of the attribute specified by domain expert and 10-fold Cross Validation yield the highest accuracy of 65.83% with precision value of 56.10% and recall value of 50% for k=3. Using weight attribute in this algorithm performs higher accuracy, precision and recall than the one which does not use it. The change in the order of importance of the attributes determined by Correlation Attribute Evaluation yield in a higher recall value of 54.35% for k=5 than the order of importance of the attributes determined by the domain expert.