Research And Applicationsin Data Mining Of The Non-taxed Revenue Audit

With the information technology widely used in the financial department,the whole process of non-taxed revenue,fund application,and accounting are under the environment of computerization.Financial department accumulates a mass of data,which contain both inventories and financial records.As data are the most direct form of expression of the information resources,auditing department which is an important supervisory department has fully realized the value of the data.As the aim of audit turning to pick out the trends and effect,improve forecasting,and make the government operate more efficiently,data analysis and processing become more and more important.Based on the mathematical model,data mining technology can be used to analyze the complex data which has been collected through long time.Data mining technology can also find the character of data,review patterns,predict behavior,improve forecasting,provide basis for policy making,evaluate the efficiency of policy,and detect problems as early as possible.By studying visualization and data mining,this thesis takes non-tax audit case as an example.We aim to explore the methods,technologies,and other content of non-tax audit with the methods of decision tree,classification,cluster analysis,and regression analysis.The traditional audit focuses on the accuracy by recalculating data from operation department and finance department.The traditional audit is facing great challenges now.The digital audit using mathematical model is different.It will pay more attention to the characteristic,including regional differences,scale variation and trend.By classification the data,auditor will find the systemic problem and risk,and truly achieve full coverage of Non-taxed Revenue Audit.