ANALISIS CREDIT SCORING MENGGUNAKAN METODE BAGGING K-NEAREST NEIGHBOR

Abstract

According to Melayu (2004) credit is all types of loans that have to be paid along with the interest by the borrower according to the agreed agreement. To keep the quality of loans and avoid financial failure of banks due to large credit risks, we need a method to identified any potentially customer’s with bad credit status, one of the methods is Credit Scoring. One of Statistical method that can predict the classification for Credit Scoring called Bagging k-Nearest Neighbor. This Method uses k-object nearest neighbor between data testing to B-bootstrap of the training dataset. This classification will use six independence variables to predict the class, these are Age, Work Year, Net Earning, Other Loan, Nominal Account and Debt Ratio. The result determine k =1 as the optimal k-value and show that Bagging k-Nearest Neighbor’s accuracy rate is 66,67%. Key word : Credit scoring, Classification, Bagging k-Nearest Neighbor