| With the popular of computer used to manage financial and business data throughout the country's administrative institutions,auditors have to overcome challenges of rapid growth of all kinds of data in their daily work.In view of the new situation,the computer audit also has to faces the challenge.How to find the meaningful doubtful audit point accurately in the massive data had become a new problem.In big data era,data mining technology is developing rapidly and the analysis method is mature day by day,which has been widely and successfully applied in other fields.The application of data mining technology in computer audit has also been reported in much research work.According to the research background,this paper summarizes the related contents of the domestic papers of the data mining application in computer audit field.The main contents and the structure of this study had been formulated.Based on the concept of computer data audit,audit object,audit method,audit platform and audit process,the basic situation of modern computer data audit had been introduced.The paper also summarized the shortcomings of the general computer audit,and puts forward the application of data mining technology to the daily on-site audit.Secondly,the concept,methods,characteristics of data mining technology and its implementation software had been discussed in detail.Because the data characteristics of data mining and audit,the paper had discussed the feasibility and necessity of applying data mining technology to the daily on-site audit work.The application of data mining model should have more objectively and effectively to explore the "abnormal" correlation in audit business,and also reduce risk of the audit.The national audit can achieve high quality and quantity.The focus of this paper is to introduce two case studies of the application of data mining technology.Firstly,the association rules are applied to hospital drug audit in order to find out whether there is an "abnormal" association between individual doctors and pharmaceutical manufacturers.Based on a new association analysis of about one million prescriptions in a hospital for a certain period of time,it is concluded that some doctors have specific "abnormal" correlations with certain manufacturers.Secondly,the Logistic regression analysis is applied to the audit of provident fund payment groups,aiming to find the possibility of high-risk loans.Based on the regression analysis of personal social security and accumulation fund information,the author diagnoses the individual accumulation fund loan,and finds the possible risk of the accumulation fund loan.Data mining technology has been applied successfully in two cases.The conclusion of this paper: data mining technology applied to computer audit is feasible.Compared to the traditional audit method,the advantage of data mining can clearly find the correlation among the statistical variables.Based on the principle of mathematical statistics,it is convenient to obtain the abnormal data and correlation performance which is not consistent with the general law.But in the actual application process,the laws and procedures of the audit process should take more consideration.Data mining technology can be used as a necessary supplement to computer audit. |
