The application of listed companies credit scoring model based on Bayes discriminant rule

Based on a brief summary of the research actualities on credit scoring of listed companies, especially the research and the application based on multivariate discriminant analysis (MDA) methods, the Bayes discriminant rule (minimum ECM rule) is introduced to establish the credit scoring model for dichotomous classification on the assumption that two groups have unequal population covariances. Then, this model is further tested and used for the pattern identification and classification of 106 listed companies in China in year 2000. At last, this article gives a discussion about the performance of this credit scoring model from two aspects: the APER and the E(AER).