An Empirical Research On Listed Company's Financial Crisis Early Warning Model

The financial crisis occurred is a gradual process, the financial crisis has not only threatened, and can be predicted. Accurately predict the financial crisis,to protect the interests of investors and creditors,on the operators to prevent financial crises. Monitoring the quality of listed companies and government departments to market risk is of great practical significance. An Empirical Study of listed companies forecast financial crisis not only has high academic value, but also of enormous social value. Based on the model of the domestic financial crisis warning area classic literature review in summary and evaluation of existing research results, based on selection of the 60 companies listed on the financial crisis and the normal 60 financial companies as a sample,we used Logistic regression, Bayesian discrimination and BP Neural Network model to establish a relatively high prediction accuracy of forecasting model. This paper is divided into three main stages: the first layer, choose 10 indicators which can significantly distinct ST companies and non-ST companies by ANOVA from profitability, cash flow and solvency,asset liability management capabilities and ability to grow a total of 22 top-five financial targets. In the second phase of research,use 40 ST companies and 40 non-ST companies at the same period as analyze samples to establish Logistic regression model, Bayesian discrimination model and BP Neural Network model. In the third stage,based on the above three models which have been established,20 ST companies and 20 non-ST companies in the same period were used as forecast sample to test,the results showed:Logistic regression's forecast accuracy,the Bayesian discrimination's prediction accuracy and BP Neural Network model's prediction all was 95% .It can be concluded that these three models have a high level of accuracy. The company's managers can choose a suitable model for specific financial crisis forecast combined with other non-quantitative factors to analyze the causes of the financial crisis so that the deteriorating financial situation can be avoided.