Commercial Bank Credit Risk Assessment Based On Rough Set

Nowadays banking still leads China's financial system. Bank's revenue mainly comes from credit business, whilst the corresponding credit risk becomes the paramount risk confronting banks. Therefore, how to control credit risk effectively and how to improve credit asset quality have been important subjects in banking industry and academic circles.The rough set theory, as one of effective methods of processing disintegrative and imprecise data, has undoubtingly played an important role in data mining. Since the time it is proposed, rough set has obtained successful application in each domain. It theoretically has profound significance. The immense potential of rough set theory attracts attentions of many experts in the world, especially its application in credit risk assessment, which is definitely a new challenge. This paper applies rough set theory into assessment of credit risk. Using both qualitative and quantitative analysis methods, the paper analyzes and demonstrates the real environment of China's commercial banks. The major achievements are as follows.The first, in the aspect of study contents and methods the paper combines rough set theory with credit risk of traditional banks, which provides a certain basis for quantitative development of classification of credit risks and reduces the subjectivity of credit classification standard, thus make credit warning system of commercial banks more comprehensive and more practically significant.The second, selecting indexes are based on both domestic and oversea achievements and operation practice of banks. The selected indexes contain financial and non-financial indicators. The whole index system includes quantitative and qualitative indicators, thus it shows quantitative and qualitative analyses are integrated in the thesis. The paper avails the strong processing function and attributes reduction function of rough set theory to select indexes from traditional financial and non-financial indexes. Reasonable deletion and addition are made on index system, which makes it more rational.The third, a credit risk assessment model is designed and implemented based on rough set theory. The process of establishing the model consists of assumptions, the choice of indicatorsdata, pre-processing, attribute reduction, risk assessment rule creation, and rule testing.Finally, the paper applies the model into enterprise credit risk assessment and individual consumption credit risk assessment. The study result shows that the credit risk assessment model based on rough set possesses excellent warning forecast capability. The total warning forecast accuracy rate are over 90%, therefore, it can be used as a reference for bank's credit policy.