Research On Control Of The Quality Of Audit Data Based On Data Mining

The changes of audit environment influence the direction of development about audit theory and practice in some ways. With the quick developing of IT and continuous enhancing of informational level, several of industries and organizations rely on the information system more and more. The auditors do not any longer face up to the paper accounting system, but the computerized information system; they do not any longer see the paper accounts, but the diversiform data. Data-based Audit emerged just in these conditions. It stands for the future of the Computer Audit in the informational circumstance with the unparalleled superiority. In the condition of information-based Audit, evidences exist in the IS of enterprises which were audited with the electronic data chiefly. It makes the quality of audit data becomes one of the most important factors that influence audit risk. Therefore, we have to see the pressure about controlling the quality of audit data. So that, we can lower the audit risk reasonably and effectively, and accomplish the audit object finally.The paper uses the standard researching method in the overall, beginning with the theory about data quality, explains the conception and features of the quality of audit data, puts forward the essentiality of controlling the quality of audit data. Whereafter, in the basic of analyzing the new characteristics about audit risk under the mode of Data-based Audit,it indicates the influences that the quality of audit data impact on audit risk, especially on the detection risk. Then the paper discusses the problems and controlling means about the quality of audit data in every phase from data gather, data collate to data analysis, from the view of technology, by analyzing the process of Data-based Audit in detail, and emphasizes the phrase of data analysis. Afterwards, it points out the feasibility that controlling the quality of audit data uses data mining, according to the demands of audit data analysis and the features of data mining. There are many kinds of measures about data mining, so in order to illustrate the standpoint, the paper takes the method of numeral analysis based on Benford's Low, and assays the application of this method for control the quality of audit data. And then, explains the problem about controlling the quality of audit data based on data mining by real data.