The Risk Pre-warning Research Of Commercial Bank That Based On SVM Method

It’s the trend of economic globalization that makes the financial sector has a stronglinkage in the entire economic system, especially the banking industry which treats thecapital as its operating targets. Once the bank credit conditions getting worse, inevitably,the liquidity problems must be followed behind and even induce the entire banking crisis,and all these will cause heavy loss to the entire economy. How to prevent the banks’ risk isbecoming a major issue of the banking industry as well as our entire country. In other words,the study of bank credit risk pre-warning has become an important research.The paper analyzes the background and significance of China’s commercial bankscredit risks firstly, and then learns from the research results at home and abroad. Besides,the paper analyzes the causes, objectivity and characteristics of banks credit risk, andimplementing the theory of bank credit risk management. Moreover, the paper analyzes themanagement model of credit risk, and expounds the existing problems, thus the paper laysthe foundation for the building of bank credit risk pre-warning model. The articleintroduces the SVM method to do the bank credit risk pre-warning research. Firstly, thepaper analyzes the feasibility of the SVM method, and on the integrated use of principalcomponent analysis and SVM to build a suitable vector machine to predict the risk whichbased on the construction of the bank’s credit risk pre-warning indicator system. The papermainly uses the financial data of publicly listed commercial banks of China to build thecredit risk pre-warning model. Firstly, the paper locks the sample data. Secondly, we usethe principal component analysis method to reduce the data dimension. And then we use theindex threshold comprehensive risk score to classify the samples into two types. Lastly, thepaper uses the common established SVM model which built in the third part to do theempirical research. The experimental results show that the SVM model has good predictiverisk. Based on this, the paper summarizes the building of credit risk pre-warning systemwhich provides recommendations to prevent and mitigate the credit risk for commercialbanks of China.