Font Size: a A A

Research On The Application Of Decision Tree Technology In The Financial Report Frauds Of Listed Companies Detecting

Posted on:2011-05-23Degree:MasterType:Thesis
Country:ChinaCandidate:Z X YaoFull Text:PDF
GTID:2189330332985191Subject:Accounting
Abstract/Summary:PDF Full Text Request
With economic development, the listed companies financial reporting fraud more and more frequent, disrupting the order of the securities market have shaken the confidence of majority of investors in the securities market, affecting the further development of the securities market. A high accuracy, fast and effective financial reporting fraud identification model, for investors to analyze financial reports of listed companies provide a reference, as much as possible to reduce the false financial report led to the decision-making errors caused by loss; also help to improve the registration of accountants financial reports of listed companies of fraud identification efficiency and reduce the risk of their audit.Papers selected from 2004 to 2008 issued the Shanghai and Shenzhen A shares listed company's annual financial report data, including data on selected financial reporting fraud from 2004 to 2008 by the SFC, the Shanghai Stock Exchange, Shenzhen Stock Exchange and the Ministry of Finance punishment 227 listed companies. After a preliminary statistical analysis, found that 227 companies of fraud involving 13 major industries, but mainly concentrated in manufacturing and information technology industry, they both accounted for 72%of the total number of fraud companies, it come to conclusion that industry is an important factors affecting the financial reporting fraud.Based on academic studies at home and abroad in the past, selected financial reports of listed companies may identify fraud indicators, the indicator system includes 11 financial indicators and seven non-financial indicators, a total of 18 indicators. Constructed the financial reporting fraud identification model by decision tree to show the form of visual tree. By analyzing the decision tree nodes found that tree down more branches, created data mining model examples of the impact split the smaller. The first branch of the tree by "accounts receivable turnover rate" property decision, which said that "accounts receivable turnover rate" is to identify the listed company's financial reporting fraud of the most important factor. Followed by a branch of "the total size of supervisors," "executives", "leverage factor" and "inventory turnover", is second in importance only "accounts receivable turnover rate" factor. These factors can serve as a financial reporting fraud identified an important signal that can be used to predict whether there has been financial reports of listed companies the possibility of fraud.Paper concludes, in order to prevent or reduce the financial reports fraud of listed companies in China must deal with the internal supervision and external supervision of two aspects:the internal supervision is to improve the governance structure of listed companies, optimizing equity structure of listed companies and improving accounting personnel professional quality and professional ethics; in the external supervision, to improve the relevant accounting standards and accounting system, improve the CPA system and to strengthen government regulation and supervision by public opinion.
Keywords/Search Tags:Detecting of Financial Report Frauds, Listed Companies, Decision Tree, Data Mining
PDF Full Text Request
Related items