Study Platform Of Analyzing Fraudulent Financial Statements Using Visualization Data Mining

The data mining technology is an interdisciplinary studies,which is rapidly develop in recent years,involves the database,statistics,artificial intelligence and the machine learning and so on many domains. The applied universality of the produced a great deal of data ,the data mining is the technique which makes use of above-mentioned science to carry on the great data quantity processing. The data mining application is extremely broad,and data mining technology will play more and more main role to future society's each domain.Fraudulent financial statements have a serious impact on capital market securities market and the investors .Preventing to issue fraudulent financial statements is a necessary and meaningful thing. Improving capability of identifying fraudulent financial statements is a effective way to solve this problem. But fraudulent means these firms use become more and more diversiform. It has become increasingly difficult to identify the fraud financial statements.As data mining technology continues to mature and become more widely applied. Data Mining Techniques become an alternative analysis for the financial statements. It can use abundant hidden information to anatomize the financial statements. We can use clustering to anatomize,compare result based on many situation and use practical meaning to explain the result. We can also use classify to analyze probability of fraudulence.Study priority of this thesis is how to structure a data mining platform,then analysis for the fraudulent financial statements using this platform. Including aspects as follows:The first,change , choose , clean , process a great quantity data with file form, and using these data build the subject database, be that the financial database lays down basis structure. The data use is download from internet,collect balance sheet,profit sheet,cash flow sheet,mid report,season report,finance summary and daily stock data,these data exist as Excel file. We build index database by VB6.0 data integration program and index compute program. We build finance fraudulent character database by collecting dada from website of SFC. Index data prepared by VB6.0 extraction program,complete the subject database and a convenience sample data extraction. The second, use visible technique to slice up multi-dimension data body, which can make users observe data from multi-angle and multi-side. There are three aspects in data mining visible technique. They are visiblizations of data, mining process and mining result. Visible data mining platform organicly integrated original data, mining process and mining result, and formed direct present of these three aspects.The third, use ActiveX groupware integrating characteristic data mining arithmetic. Systemic data mining module in this paper adopted groupware form, each data mining arithmetic was built as Active X, which could embed data mining module conveniently. Sort arithmetic illustrated by BP nerve net. Characteristic arithmetic illustrated by main element analysis. Data mining includes character chosen, sort, assembling and related groupware.The fourth, make static, trend and same trade analysis by this data exhibition module on financial data and index. Data exhibition module presented data as three methods which could analyze index quantity relatives, trends and distribution in business on a certain time in the company. Static analysis aimed at financial data and index of a company during a period. This paper analyzed four breweries structure of current asset, capital and debt, then, primarily estimated these four companies basic status. Trend analysis aimed at financial data and index of listed company in different period. This paper analyzed currency capital, payable account and the trend of stock, got the financial fraudulent phenomenon existing in the company. The conclusion is that, if management was in healthy process in a company, the financial data should be in continual and steady improvement. If exist big fluctuation or deviation between financial data and index, it means that there were fraudulent in the company. Analysis in the same trade is to find serious deviation between financial data and index. This paper analyzed same four breweries and tried to find same characters. We could know that the improvement of a listed company must be affected by macro economy operation and industry state. Even the primary enterprise will not be badly depart the average level. So if a company's financial status seriously departed industrial average level, it should be some problems.The last, is to make selection for financial fraudulent characters and sort companies out by this module and nerve net. This paper adopted two modules to analyze financial report forms and got valuable reference data. Different data nerve net sorts were used in the paper. One is the random sample extracted, and the other is selective extracted according to collecting result. Demonstration analyzed that the latter method had strong ability to detect sealed module of data. Especially to identify financial fraudulent practices, and collecting method also could provide data selecting reference to improve sort module and advance veracity of identification. By this characteristic data, the writer got final sample to confirm whether the similar record represent usable things in reality.This paper analyzed characters of financial fraudulent behavior in listed company used data mining platform. By using nerve net sort to achieve sample data practicing, get testing data and improve identity of financial fraudulent behavior.