Research On Financial Fraud Information Identification Of Chinese Pharmaceutical Listed Companies

Financial information,as an important information resourcereflects the financial situation and operating results of listed companies,plays an important role in business management and investment decision-making.However,modified financial statements will not only mislead investors and management decision-makers,but also lose the function of risk warning,increasing the operational risk of the market.Even the capital market,information disclosure,financial supervision,corporate governance and organizational structure in the United State are well-established,but there also exist the financial fraud between the listed companies.In China,with the development of market economy is becoming more and more mature,the investor group is increasing,and the risk awareness of investors in the process of investment is gradually increasing.Therefore,the financial fraud problem of listed companies has been paid more and more attention.Every year,some listed companies in our country have been investigated and penalized by the CSRC for financial fraud or because of financial problems.However,the financial fraud of listed companies is still a problem.At present,it is of great significance to search for a set of accurate financial fraud recognition model to discover the financial fraud information of listed companies.This article first puts forward the background and significance of the research,summarizes the research status at home and abroad,clarifies the main contents,main methods and the structure of the article;and then defines the connotation of financial fraud.Furthermore,the methods of financial fraud identification: PCA-Thompson tau combination method,decision tree method,Logistic regression method,support vector machine method(SVM)and other methods were introduced.On this basis,this paper chooses three representative companies from the listed companies of fraudulent pharmaceutical companies as the object of study.Firstly,the differences of the primary sample data are tested by significance test and correlation test.From the initial 26 financial indicators,the paper chooses the final 10 financial indicators to conduct the research.Based on the selected financial indicators,the paper uses PCA-Thompson tau combination method,decision tree method,Logistic regression method and support vector machine method(SVM)to recognise and forecast the financial fraud,finally summarized the research results of the financial fraud recognition,and pointed out the research existence limitation and the follow-up research direction.The results show that,before modeling analysis,it is helpful to simplify the work of modeling and improve the correctness of the model by eliminating the different data,testing the original data and eliminating the redundant data.At the same time,the precision of financial fraud recognition and forecasting model ofthe listed companies analysied based on SVM method is higher than that of PCA-Thompson tau combination method,decision tree method and Logistic regression method.It can help government regulators,investors and auditing departments to correctly identify the financial fraudment of listed companies.