The Development And Research Of Audit Sampling Systems On The Basis Of Data Mining Technology

With the increasing popularity of computer technique, using it to substitute manual accounting has many advantages. Computerized accounting and E-business data are widely used by many companies. Audit sampling of financial and business data is the basis of daily audit for audited units. How to find the potential useful information in these large and chaotic mass data is of vital importance. So, the auditors are eager to find out things of real value in these mass data in order to provide the clues and bases for solving problems.This paper explicitly describes an audit sampling system which the author participated in and developed. Data mining technique upgrade the stratified sampling technique to a new area. This paper includes:1. It interprets the concepts of data mining and its current situation as well as some relevant content of audit sampling both nationally and internationally. It also analyses the significance of data mining technique in audit sampling system and introduces techniques related to clustering and association rules so as to reveal the process and significance of data mining in audit sampling.2. An improved SRISP-DM data mining model is set through the analyses of data mining process and its actual application of audit sampling. And then, to design this audit sampling by the improved MVC model of AJAX technique which utilizes EXTJS framework to create a front-end user interface independent of back-end technique. Because using EXTJS is more convenient and quicker in finishing tasks at view layer and using AJAX can reduce the redundant requests and lighten the servers' load for its advantages, such as limiting the waiting time, alleviating the users' pressure and no need of refreshing the interface.3. This paper proposes an improved K-means algorithm based on optimization of a traditional stratified sampling method under the premise of guaranteeing the sample data as typical one. This not only reduces the number of audit sampling, but also improves its accuracy. Using Aprioir's association analysis to do association analysis on audited units, finding economic responsibility audit and high risk audit of fixed asset investment will provide useful information for auditors in choosing auditing object and improve the auditing quality.