Detection Research Of Financial Irregularity Of Chinese Listed Company Based On Naive Bayes Classification Algorithm

In recent years,with the development of information technology,corporate finance continue to carry out transformation and upgrading,the company's financial work from the initial manual accounting,to the development of information technology,and accounting for financial management to financial changes,and then shift to strategic finance,Corporate accounting methods and financial thinking of the ever-changing,corporate information disclosure is more and more transparent.But in any case change,the financial situation of the enterprise has always been subject to regulatory authorities,government agencies,external investors,the concern of the internal business operators,corporate financial situation of the small abnormalities for the enterprise accounting assumptions of continuous operation,monetary measurement will bring Significant impact,and may be associated with the parties to bring huge losses,so the study of corporate financial anomalies is very important.There are many reasons for the financial abnormalities of enterprises,some are because of having problems in the business,some are enterprises for specific purposes for financial fraud,but there also are caused by the enterprise's internal control system design is not enough to increase the risk of financial abnormalities.But for whatever reason,the company's financial anomalies will eventually be reflected in the fluctuations in financial indicators.Therefore,this paper chooses the financial indicators of Shanghai and Shenzhen A-share listed companies from 2003 to 2015 due to abnormal financial situation as a sample.Using the significance test in SPSS,it is judged whether there are significant differences between financial abnormalities and non-abnormal enterprises.There are 32 indicators of significant differences between financial abnormalities and non-abnormal enterprises.By using the factor analysis extracts the indicators of the following listed financial indicators,extracting the main role of the financial indicators of the formation of 11 new financial indicators for the next study.SPSS is used to discretize the data,and the continuous index data is converted to discretized data using equal percentages based on scanning cases.By using the method of Naive Bayesian classification,the probability value of the characteristic attribute of the classifier is determined by using the training sample,and then the accuracy of the Bayesian classifier is verified by the processing of the test sample.The Naive Bayesian model the overall accuracy rate was 77.92%,the error rate was 22.08%,the rejection rate was 12.34%,and the error rate was 9.7%.In this paper,we use the method of naive Bayesian to study the financial abnormalities of listed companies.It is found that the financial abnormalities of enterprises are affected by industry.The finance of manufacturing industry and real estate industry is more likely to be abnormal.In recent years,the development of more rapid information Technical,software and information technology companies are relatively more financial status anomalies.The financial performance of the enterprise's financial indicators,property rights structure indicators of the enterprise's financial anomalies have a strong indicator,the majority of the financial situation of enterprises there are consecutive years of abnormal.