Segmentasi Nasabah dalam Pengembalian Kredit dengan Metode CHAID

Abstract

Credit is one part of capital formation carried out by financial institutions, in this case the banks to public. Credit risk is the potential loss of consumer credit refusal or inability to pay its debts in full and on time. Customers in the level of credit risk is influenced by some variables. But in reality, too many variables that would complicate the determination level of credit risk for the bank in the future. Because of this, it is necessary to simplify the variables in a way to sort out which variables are the most significant influence. CHAID (Chi-Square Automatic Interaction Detection) method is one of method can be applied to address this problem. Results of analysis of this study is that there are four explanatory variables that are associated with the structural status of the collectability of customer. The variables are the number of installments, job, loan term, and gender. CHAID analysis produces six customer segments. From the table, it can be concluded that the classification accuracy of prediction for the current customers by 96.2% while the predictive accuracy of customer loss amounting 10.0%.