Font Size: a A A

Weighted A Posteriori Support Vector And Its Listed Companies' Financial Credit Assessment

Posted on:2012-05-18Degree:MasterType:Thesis
Country:ChinaCandidate:J P TengFull Text:PDF
GTID:2199330335989642Subject:Probability theory and mathematical statistics
Abstract/Summary:PDF Full Text Request
With the intensifying of economic reform in our country, enterprises are facing with the fiercer and fiercer market competition. In Chinese stock market which is growing with every passing day but immature, the financial condition of some listed companies gets into trouble frequently. So to establish practical and effective credit evaluation model for credit assessment of listed companies has become a financial necessity and is significant for the development of China's securities market.This paper is centred on four issues:First, with the beginning of analysing the necessity of credit evaluation for the corporate, this article analyse the actual situation of the credit evaluation industry combining the credit status of our society. Then the key performance indicators and evaluation method for quantitative test to corporate credit conditions was designed by using quantitative methods. Second, during the process of selecting sample data, we make use of the stratified sampling first, systematic sampling later to ensure representativeness of the sample. After selecting sample data which reflect the financial index of the listed companies, they should be filtered by the method of variance analys. Third, at first we can get the principal component factor score values through principal component analysis,then select the training samples and testing samples, after that train the sample by using PPSVM-weighted to established the model, finally, check the test samples in order to get classification accuracy and error of the training samples and testing samples. In the end,by comparision of train accuracy and test error with the SVM or the other credit evaluation method,it illustrate the advantage of PPSVM-weighted method. PPSVM-weighted theory is put forward.The combination forecasting model based on principal component analysis and PPSVM-weighted theory is established.This method is not only suitable for the sample we select but also for the other companies. The assessment of credit level for the enterprises according to the fianancial data index of any corporates can be applied in government regulators,banks and largeinvestors for the decision making analysis. Besides,it palys an early warning role for financial credit of the corporate.
Keywords/Search Tags:principal component analysis, support vector machine, credit evaluation, financial indicators
PDF Full Text Request
Related items