Bayesian credit scoring model with integration of expert knowledge and customer data
24th Mini EURO Conference on Continuous Optimization and Information-Based Technologies in the Financial Sector, MEC EurOPT 2010, İzmir, Türkiye, 23 - 26 Haziran 2010, ss.324-329, (Tam Metin Bildiri)
- Yayın Türü: Bildiri / Tam Metin Bildiri
- Basıldığı Şehir: İzmir
- Basıldığı Ülke: Türkiye
- Sayfa Sayıları: ss.324-329
- Anahtar Kelimeler: Bayesian classification, Credit scoring, Expectation-maximization algorithm
- Boğaziçi Üniversitesi Adresli: Evet
Özet
Bayesian classification method is one of the effective classification methods in credit scoring applications. Application of this method to credit scoring provides several advantages, which are suggested in the literature. In this research, a Bayesian credit scoring model is applied for classification of personal loan customers. The contribution of this paper is two-fold. First, we apply a Bayesian classification model to credit scoring of personal loan customers with integration of knowledge of credit experts and personal loan data. The dependency relations between the variables are constructed based on the knowledge of credit experts, and the parameters are estimated by Expectation-Maximization (EM) algorithm with real personal loan data of customers obtained from a Turkish bank. Second, we assess the effectiveness of the Bayesian credit scoring model based on additional personal loan data of the bank customers. The results show that the Bayesian credit scoring model is more effective than the logistic regression credit scoring model which has been currently used in the bank. © Izmir University of Economics, Turkey, 2010.