The model and application of mining the features of financial statement audit with multi-subjects

When auditing the large enterprise groups with many accounting subjects, in order to find. the doubtful auditing points quickly in the mass electronic data, we designed and developed a accounting report procedure for single subject according to the logic of balance sheet with the parallel simulation; adopted the associative rules to mine the audit features of electronic data, and combined with the auditing experience to analyze the abnormal data, abstracting the features of fraud. Finally, took an example to test the effectiveness of the model. We quickly found the audit trails of many companies, which reduced the risk of incorrect acceptance caused by the delay of data acquisition.