ABSTRACT
Credit risk refers to the risk to the asset income of banking credit due to consumer’s personal failure of fulfilling the contract signed with bank. The best way to minimize the loss of credit risk is to set up a reasonable credit risk evaluation model[1]. At present, most credit risk evaluations mainly adopt qualitative method with only two outcomes of “Good Credit” or “Bad Credit”. In fact, the credit risk evaluation needs quantitative evaluation, e.g. the degree of “good” or “bad” credit. On the basis of attribute theory, the attribute coordinate evaluation is a comprehensive evaluation method that realizes multi-agent evaluation[2] and reflects the evaluator’s preference curve with both qualitative and quantitative analysis. This paper conducts risk analysis of the attribute characteristics of creditors and their credit status from the German Credit Data in the UCI machine learning database[3].
