Logistic Regression And Its Application In Credit Risk Measurement Of Listed Companies

With the development of China’s economy,credit plays a more and more important role in the economic system.The increasing credit transaction has led to the increase of credit risk,and more transactions have made credit more and more widely used in life.The increase in the use of stocks and bonds in recent years has increased credit risk,and the incidents of default have emerged,making the public a strong demand for the accurate and timely measurement of the credit risk measurement model.As the main body of stock and bond issue,the listed companies have huge volume and are most closely related to the public.Because of its important position in the national economy,it is important to find a reasonable model to evaluate the credit risk of the listed companies and to protect the interests of the investors.Through the analysis of the current situation of China’s companies,it is found that the method of measuring the credit risk of the company is lagging behind the developed countries,and there are also defects in the selection of data and the treatment of the corresponding proportion.However,due to the different development of our country,it is not advisable to copy foreign experience completely.By analyzing the situation of our country’s development and the analysis of information data acquisition ability of each company,combining the present situation of the credit risk model and the advantages and disadvantages of each model,this paper decides to use the Logistic model to measure the credit risk of the listed companies.In this paper,the proportion of 1:3’s ST and non ST company is selected to ensure the agreement with the reality.A total of 268 companies are selected and randomly divided into training sets and test sets.Considering the availability,authenticity and effectiveness of the data,a series of financial indicators are selected and the total asset indicators are added to make the model even more.All round.For the selected 31 indexes,22 indexes were screened out by Kolmogorov-Smirnov normality test,independent sample t test and Mann-Whitney U test.Then KMO and Bartlett spherical tests were used to carry out the principal component analysis to reduce the dimension,and 8 main components were selected.The Logistic regression model was constructed and tested with the index.Finally,combined with the results of this study,the role and applicability of credit risk model measurement are analyzed.