Research On Credit Risk Assessment Of Personal Loan Business Of Guizhou Bank

With the transformation of Chinese personal consumption concept,the demand for personal loans is increasing,too.In particular,the rapid development of Chinese real estate industry in the past decade has greatly promoted the rapid development of personal loans demand.But the US sub prime loan crisis in 2008 was a wake-up call to the world,and Banks began to pay more attention to the management of credit risk of personal loans.Around 2013,all banks in China case prevention situation was severe extremely.From the Qilu Bill case,the Qingdao port fraud case,to the national steel trade credit crisis,major cases in various regions frequent,a large number of bank capital has lose.Thus,the CBRC issued a notice on strengthening efforts to prevent operational risks(article 13),requiring Banks to take measures to prevent and control operational risks.In 2004,Basel II defined the major risks of commercial Banks as credit risk,market risk and operational risk.And credit risk is the most important and complicated risk that commercial bank faces in its business.Guizhou Bank has accumulated certain historical data in the operation of the past six years.How to use of these data rationally by scientific analysis,improve service quality and prevent market risks is of great significance for the bank in its future development.Discriminant analysis is known as Linear Discriminant Analysis,which was born in the 1930 s.It is a statistical method based on historical data,through the analysis of the historical data of sample type and the establishment of discriminant model,then using the discriminant model to classify the unknown types samples.Since the late 1990 s,discriminant analysis has been widely used in disease prevention,plant and animal species selection and economic management.Discriminant analysis has distinctive features,that is analyzing the data of several samples of each known category and summarizing the regularity of classification of objective things so as to construct the discriminant formula and discriminant criterion.When the new sample point of the event is encountered in the future,the type of the sample point can be determined with high probability only according to the constructed discriminant formula and discriminant criterion.This paper found that the person age,income level,education background,job category,loan to value ratio(LTV),loan orientation and so on factors will affect the personal loan default by classifying the past personal credit data statistical analysis,which is based on the historical data of Guizhou Bank database.After quantifying the data,the paper uses the factor analysis to reduce the data dimension,firstly,and then using the Fisher discriminant analysis and Logistic regression model to construct the discriminant function,through the empirical analysis we found that the accuracy of two methods is more than 80%,finally,as a comparison,K-means clustering analysis was added to,and we found that the discriminant analysis was superior to clustering analysis.