The private credit risk prediction represents an assessment of the relative level of risk associated with the timely payments required by the debt obligation. Credit scoring system that is most common screening method created for the evaluation of credit card customer is based on the available statistical information that is related to the behavior of former clients with credit.
There are different characteristics in the field of credit scoring: (1) Continuous target variable, (2) Lots of case bases, and (3) high correlation between target variable and independent variables. This paper suggests that for the characteristics of the field, Case-based reasoning will fit in the credit scoring problem. We will use three different techniques: (1) Case-based reasoning, (2) Neural network, and (3) Statistical methods to predict credit card customer credit and to evaluate the prediction ability of the techniques.
Generally neural network is famous for its highly predictable resuluts. But, it also has a weakness that the results can not be explained because of its black box approach.
But, opposite for that, Case-based reasoning(CBR) is famous for its readily understandable results. In the field of credit scoring, forecasting credit scores only is not enough, because after that there must be an analysis of the result that the certain technique made.
In this point of view, if the results from CBR is better, or even a little inferior to other two techniques, we can suggest that CBR is a proper technique for the credit scoring area.