With the development of science and technology, the relationships among human beings especially their economic relations turn to the complex form, so economy and business entered a new era. It is difficult for investors to make decision because of the fluctuations of stock therefore they need to predict the bankruptcy. It has been a long time that discussion about the models of bankruptcy or financial crisis in accounting and management has been under consideration. Among the methods that can be used to predict bankruptcy, the methods that use financial ratio analysis is more valid. The purpose of this study is adjusting the models of bankruptcy prediction by using comparatives models such as logistic regression, discriminate analysis, and data envelopment analysis(DEA) methods for predicting the bankruptcy status ofcompanies regarding the environmental requirements of Chinese companies in shanghai stock exchanges and Shenzhen stock exchanges as well as. A Bankruptcy laws were later issued in 1915 and 1935 during the Republic of China period, which ended in 1949 when the People's Republic of China was founded. Some notable bankruptcy cases like Sanlu in November 2008, East Star Airlines in early 2009, Taizinai Group in 2010, Jinhua Chemical Industry in 2010, Life Show in 2012, Xin Dongxiang Business Hotel in 2013 and so on were happened in China that is make stockholders and investors to think about bankruptcy more. In this survey we analyses the most famous bankruptcy models such as William Beaver model, Altman models(Z, Z', Z”), Springate model, Ohlson model, CA-SCORE model, Yushiko Shirata's model, Zmijewski model, Fulmer model.For analysis the first hypothesis which is adjusting the previous models by Logistic Regression, first we using Omnibus Test for testing the efficiency of the models, second Cox & Snell R Square statistic are circumstantial evidence that determine how much independent variables extend and explaining the variance. After that we used Hosmer and Lemeshow statistic for goodness of variables. The last we used the Wald statistic and we introduce the success of the adjusting models by Logistic Regression model to predict the bankruptcy of companies.The second assumption is about adjusting the previous models by Discriminant Analysis, so we use the Box's M test to check the equality of covariance. For testing equality independent variables of several groups we used Wilks Lambda Test. Also for calculating diagnostic equation the correlation between each independent variable with diagnostic function should calculated, and then we have the Success Discriminant Analysis Method after all our analyses.For testing the third hypothesis we choose the total assets, total liabilities, and sale expenses as input variables, and income as output variable. Also we used DEAP software for analyses.Another goal of our research is finding an ideal bankruptcy model with a lower standard error for companies between previous models, According to the analysis, the best model for predicting bankruptcy around Chinese exchange company is Springate model and the worst model is Yushiko Shirata, and other models order by priority are as follow:Springate, Z-Altman Model, Z”-Altman Model, Z'-Altman Model, CA-Score, Fulmer, Ohlson, Zmijewski, Yushiko Shirata.Furthermore for the last aim of this survey we find the new model for predicting the bankruptcy by applying the data envelopment analysis(DEA) regarding the environmental requirements of companies by using financial ratios, so first we test the liner relationship between our financial ratios. For checking liner relationship to get out target we use Tolerance and Variance Inflation Factor(VIF) statistics. According to the analyses, among the 32 financial ratios just X2(Net Income Ratios), X3(Total Debt to Total Assets), X6(Retained Earnings / Total Assets), X7(EBIT / Total Assets), X9(Sales/ Total Assets), X21(Current Ratio), X22(Book Value of Equity / Book Value of Total Liabilities) and X23((Current Assets- Current Liabilities) / Total Assets) have no liner relationship. after all processing element the new model where created with ija are that Indicate input variable(X23, X3, X22 and X21) and rjbare that Indicate output variable(X6, X9, X7 and X2). By using input and out data we found the new model for predicting bankruptcy by more than 94.6 present accuracy. |