Financial Crisis In The Logistic Regression Model's Construction And Testing

Since 2007, the U.S. subprime mortgage crisis has triggered a worldwide economic crisis and spreaded further,It has endangered and overwhelmed many companies, so many comies have closed down. become all country, and all enterprises increasingly attach great importance to Financial early warning . This study aims to analyze and screen through the testing of the financial indicators listed companies,In order to find the financial crisis could significantly differentiate from non-financial crisis in the company's financial early warning indicators, and establish China's financial early warning model of listed companies through Logistic regression model. I hope to help listed companies discover and resolve the financial crisis that will occurin a timely manner, so the company take to the track of normal development.This paper has been written drawing on the basis of previous studies and had been used theoretical analysis and empirical research method,.Statistical methods used in empirical research is SPSS statistical software, Sample was 465 listed companies manufacturing samples and test samples into training, the following major elements:Chapter 1 Introduction.This chapter sets out the background and significance of thesis topics, and the methodology of the research methods and ideas.Chapter 2Financial Crisis Summary of Research Documents and functions. Detailed review of the business services and foreign scholars on early warning of the main research results, and the Financial Crisis of the meaning and function are discussed.Chapter 3 the sample, variables and model choice. Mainly selected according to certain standards of construction Logistic regression model of the sample group and duration, the research since 2006 to the year 2008 China's Shenzhen and Shanghai's 76 ST and non-ST of 399 manufacturing enterprises, selected 22 a traditional financial indicators and the six indicators of cash flow modeling. Finally the three key financial early warning Models, Logistic model of the decision to establish this model of financial early warning.Chapter 4 the Logistic Model Pre- Warning construction and the applicability of tests. Financial indicators on the basis of non-parametric tests, principal component analysis, and gradually remove variables and other methods, Select a significant contribution to the model of financial indicators into the model, In accordance with the first three years of financial crisis, were constructed for three different years of financial distress prediction model. The results showed that the financial crisis before the T-1, T-2, T-3, respectively, on the validity of test: 90.51%, 86.78%, 74.03%. This fully illustrate the three years before the establishment of three early-warning model achieved good prediction results have great practical value.Chapter 5 Conclusion. In this paper, the conclusions drawn, this study characteristics, and follow-up research prospects.