Data Mining Technology In The Audit

With the application of computer technology in auditing, the thinking model of auditing has changed largely. It is vital in auditing work whether the method of analysis is right or advanced. The main trend is that auditors are requested to master the relationship between the whole and the part by using systematic thinking and the overall or perspective point, for narrow judgment has become a thing of the past.Today, the method of analysis in auditing work is from the provision of laws and rules, business processing logic of audited institution, relation of the auditing data, mapping of inner and outer data, and experience from previous work. However, another way of carrying out auditing work is by using data mining, a data analysis tool which can help auditors find potential relationship and rules from mass data.Presently, loan risk of our nation's business bank is classified into five categories: normal, concerned, secondary concerned, doubtful, and losing. The first two are normal and the remaining belong to unsound loan. For certain reasons, some unsound loans may wrongly be regarded as normal so that the true condition of loan risk is covered up. Therefore auditors should choose some main loaning units to extend auditing work from normal and concerned loans.The paper set up models for the loan risk of a business bank by using two decision-making tree based on Clementine. We studied the sample data by means of data analysis and statistics. After analysis we found some laws. During the experiment, the correctness of the model based on C5.0 is 77.79% and the correctness of the one based on CART is 76.9%. Both results are satisfying and so is applicable to loan risk predicting analysis for business banks.