Research On The Early Warning Mechanism Of Personal Credit Credit Risk Of China 's Commercial Banks Based On Support Vector Machine Model

With the rapid development of China’s economy, the individual credit risk that commercial banks are facing is becoming increasingly difficult to manage and control. In this situation, to build an accurate model for warning mechanism of individual credit risk to assess the credit risk of individual credit customer is significantly important to China’s commercial banks.Basing on the results of previous studies and drawing on the results of enterprise credit risk research on commercial banks, this paper proposes to build an accurate model for warning mechanism of individual credit risk and consists of four parts. First, it introduces relevant concepts and theories of credit and warning. Second, it introduces the research situation of risk warning mechanism of credit risk of China’s commercial and points out the urgency of building credit risk warning mechanism of China’s commercial banks. Third, research about how to build this warning mechanism is given. It comprises of three parts which are setting of warning level, building of warning indicator system and selection of warning model. Among them, the setting of warning is made according to the actual situation of domestic individual credit risk and other financially similar countries. The building of indicator system partially refer to general standard of domestic commercial banks and through analysis has been done to every indicator which includes basic situation, working conditions, repayment ability and repayment willingness. The selection of warning model is made according to actual situation of domestic individual credit and support vector machine is chosen as theory model of this article. Last, with the data collected from 358 valid questionnaires for credit customers of commercial banks, training and testing samples are selected to reduce the dimension of 19 index adopting principal component analysis first and then support vector machine model used to assess with a accuracy percentage of 87%. At the same time, a case analysis is picked from a Industrial and Commercial Bank of China in Hefei with results indicating accuracy percentage of 92.9% which indicates that support vector is an effective and accurate way to assess individual credit risk.Through the research we find that the model of personal credit risk warning mechanism for commercial bank in this paper is an effective warning mechanism of personal credit risk. It can predict well the credit risk, control and reduce loss. It also provides a classification methods to quantifiably support the credit departments of finance institutions, as well as helping China’s commercial banks and other financial institutions to efficiently monitor and manage individual credit risk, improve the level of credit risk management and reduce losses.