This article focuses on the research of the c orre ctne s s- di stingui shing o f datasemotme problem thh is falen across in the transaction-based auditing and thefuzzy synthetic evaluation problem that is faced in the system-based auditing.Author gives the theory and the method of imPlemelltaion aPplying expertsystem to correctne s s - di stingui shing o f d ata sem ameme on the basi s o f the theo ryof c orrecine s s -di stingui shing o f data semaneme on re lational mo del. Aiming atthe sPecialty of the distinguishing of data semaneme problem, it put forWars theAnoWedge exPression method based on classified production rules and thecorresponding matching algoritluns and reasoning method. The main problem ofthe auditing of internal control system (ICS) is unaccurae. By aPplying theimProved method of fuzZy auditing, the accuracy of the evaluation and auditing isheightened. For siInPle ICS, the fuzZy auditing-fuzzy synthetic evaluaion andauditing (FSEA) is suitable. For complicated ICS, it gives the multistage FSEAmethod.According to above-mentioned result of theory research, the aUthor developsthe comPuter aiding auditing system on the basis of the distinguishing of datasemaneme and the fuzzy synthetic evaluation. By practical experimellt, it gets theconclusion that it is feasible and reliable to distinguish data semaneme ofdatabase by artificial intel1igence and database method; for comPlicated ICS, theperformance ofmultistage FSEA is bcher than FSEA.
|