Research On Financial Crisis Of Listed Companies In Manufacturing Industry Based On Logistic Regression And SVM

At present,in the context of the fast-growing Chinese economy and fierce corporate competition,the number of companies experiencing financial crisis is rapidly increasing,so the prediction of their financial crisis is very important.The productivity level of a country is closely related to the level of development of its manufacturing industry,and manufacturing accounts for a large part of the national economy of developed countries.Manufacturing is an important industry and foundation for the sustainable development of the national economy.It is a very important material cornerstone for the development of modern civilization and an important guarantee for modern competition,national defense equipment and national security success.Therefore,how to combine the actual situation of listed companies in China's manufacturing industry to construct a good model to predict financial crisis has become a more important issue.This paper selects the data of 399 listed companies in 2017 and analyzes them.First,factor analysis is performed to reduce the dimensionality of the data,and then the results of the factor analysis are analyzed.Secondly,the results of the factor analysis are applied to the logistic regression model and the support vector machine to construct a logistic regression model.Finally,the results of the factor analysis are also Applying to SVM,the initial SVM model and the SVM model after five-fold cross-validation are constructed and compared with the previously established logistic regression model to analyze the accuracy of model prediction.The analysis results show that the accuracy of the initial SVM model and the cross-validated SVM model is almost the same as that of the logistic regression prediction model.2.The SVM model with RBF kernel function obtains the best effect through the initial SVM model,but because The accuracy of the prediction of financial crisis in the test set is very low,so factor analysis is necessary.3.In the prediction of financial crisis companies,the accuracy of the SVM model is much higher than the logistic regression model and the initial SVM model,reaching 92.6%.Therefore,the model after cross-validation works best,and the accuracy of the test set and training set reaches 99.5%.Combined with the research conclusions of this paper,suggestions for the financial crisis forecast of listed companies in China's manufacturing industry are put forward.