Credit Risk Evaluation Using SVM with Direct Search for Parameter Selection
With the rapid growth and increased competition in credit industry, credit risk evaluation is becoming more important for credit-granting institutions. A good credit risk evaluation tool can help them to grant credit to more creditworthy applicants thus i
- PDF / 688,710 Bytes
- 15 Pages / 595.276 x 841.89 pts (A4) Page_size
- 52 Downloads / 196 Views
3.1 Introduction With the rapid growth and increased competition in credit industry, credit risk evaluation is becoming more important for credit-granting institutions. A good credit risk evaluation tool can help them to grant credit to more creditworthy applicants thus increasing profits. Moreover, it can deny credit for the noncreditworthy applicants and thus decreasing losses. Currently the credit-granting institutions are paying much more attention to develop efficient and sophisticated tools to evaluate and control credit risks, which can help them to win more market shares without taking too much risk. In recent two decades, credit scoring is becoming one of the primary methods to develop a credit risk assessment tool. Credit scoring is a method to evaluate the credit risk of loan applicants with their corresponding credit score which is obtained from a credit scoring model. A credit score is a number that can represent the creditworthiness of an applicant and it is based on the analysis of an applicant’s characteristics from the application file using the credit score model. The credit-granting institutions can use the scores to categorize credit applicants as either a “good credit” group whose credit scores are beyond a threshold value and so should be granted credit or a “bad credit” group whose credit scores are below a threshold value, therefore, should be denied credit. The credit scoring model is developed on the basis of historical data about the performance of previously made loans with some quantitative techniques, such as statistics, mathematical programming, data mining. A well-designed model should have higher classification accuracy to classify the new applicants or existing customers as good or bad. In order to obtain a satisfactory credit scoring model to minimize the credit losses, a variety of credit scoring techniques have been developed. First of all, most statistical and optimization techniques has been widely applied to build the credit scoring model such as linear discriminant analy-
42
3 Credit Risk Evaluation Using SVM with Direct Search
sis (Myers and Forgy, 1963), logistic regression (Wiginton, 1980), knearest neighbor (KNN) (Henley and Hand, 1997), decision tree (Makowski, 1985), linear programming (Hardy and Adrian, 1985; Glover, 1990), integer programming (Gehrlei and Wagner, 1997). Although these methods are relatively simple and explainable, the ability to discriminate good customers from bad ones is still an argumentative problem. Second, some new emerging artificial intelligence methods have also been used for developing the credit scoring models, such as artificial neural networks (ANN) (West, 2000; Jensen, 1992), genetic algorithm (GA) (Yobas et al.,2000; Desai et al., 1997) and support vector machines (Van Gestel et al., 2003; Schebesch and Stecking, 2005; Lai et al., 2006c). Some comprehensive introductions about the methods in credit scoring can be referred to the three recent surveys (Baesens et al., 2003; Rosenberg and Gleit, 1994; Thomas, 2000). The support vector
Data Loading...
