Early Warning Of Financial Crisis Of Listed Companies Based On PCA-w-SVM

Financial crisis early warning is a worldwide problem.In the production and operation process of an enterprise,risks exist all the time,but the enterprise usually controls the risk within a certain level and maintains daily operations on this basis.If the risk is not controlled and continues to accumulate,it will become a financial crisis.It does not happen instantaneously,financial crisis can be identified before the company is sentenced to financial failure.If the signs of financial crisis are identified,stakeholders will take timely action before the financial crisis really arrives.This article focuses on the issue of early warning of financial crisis of listed companies.It selects the Chinese listed companies from 2010 to 2020 as the research sample,uses ST as the criterion for judging the financial crisis of listed companies,and introduces principal component analysis and weighted support vector machine model,in order to obtain an effective method for early warning of financial crisis.The research of this article is divided into two parts.The first part is to use qualitative and quantitative research to construct the financial crisis early warning indicator system.First of all,this article refers to related theories and existing research results,and obtains an indicator system for financial crisis early warning through qualitative research methods.The indicator system includes three categories: financial indicators,corporate governance indicators,and market performance indicators,with a total of 37 three categories level indicators.Next,from a quantitative perspective,this article uses correlation test and principal component analysis to screen the indicators.These principal component factors contain most of the information of the original index system.These 9 principal components factor will be used as the feature vector of the input weighted support vector machine.The second part is to use the weighted support vector machine model to predict the financial crisis,and introduce a confusion matrix to evaluate the performance of the model.First of all,this paper constructs a weighted support vector machine model.By learning the samples from 2010 to 2019,adjusting the optimized parameters w to make the model show the best performance,using the 2020 samples as the test set samples to test the performance of the model,and then this article Using confusion matrix to analyze the model,the analysis found that the overall accuracy of the model exceeds 90%,and the ability to identify financial crisis companies is strong.Therefore,the PCA-wSVM model can effectively predict the financial crisis of listed companies.This model is a financial crisis early warning model suitable for Chinese listed companies.