Comparative Study On Models Of Financial Distress Prediction Based On Chinese Listed Companies

Along with the development of the global economic, the competition in the market isbecoming intense. A company would encounters many risks and uncertainty in itsmanagement. How to guard aginst the financial risk and the crisis is the question that eachenterprise must faces. Therefore, financial distress predicting for the companies becomesan important research topic in the field of modern enterprise finacial management. Theoccurrence of the financial distress is a gradual process, and it not only has the aura butalso can be predicted. The company fall into a financial distress, not only crisis its owndevelopment, but also bring the massive loss to the investor and the creditor. Therefore,the foundation of financial distress pre-warning system, makes the forecast to the financialoperation, is good for the company, investor, creditor, bank and so on.Since1960s, more and more companies bankrupted. The researchers in manycountrys try to predict bankrupt through many kinds of ways. In the recent50years, manymodels come out, such as multivariate discriminant analysis, logistic regression analysisand neural network model. Because of the different financial indicators, researchers getdifferent conclusions. It’s hard to say which model is better than others. We choose thesame sample data, use the multivariate discriminant analysis model, logistic regressionmodel and BP neural network model to predict the companies’ financial distress, then wemake a comparative analysis, and get the most precise predicting model.we creat the predicting models and make the empirical study based on the threepredicting models and the finance data which collects, and then compare the accuracy. TheEmpirical study’s result indicated that BP neural network model is the most pricisepredicting model in this article, and then the logistic regression analysis, the lastmultivariate discriminant analysis. But all the three predictive method’s accuracy areabove80percent, this means all the three models are effective in the financial distresspredicting. They can predict wether a company could fall into a financial distress.