关键词: 关联规则挖掘, 自组织神经网络, 审计线索

Abstract: In order to achieve the audit trail of the massive data quickly found through data mining FMA algorithms to quickly extract trial data and audit expertise library association rules；re-use of self-organizing neural network improved CLARANS algorithm to extract audit expertise library divide a similar rule base rules；then by trial set of association rules and expert experience similar rules group relative strength, the approach value and the different rate of comparing the resulting set of audit trail.

Key words: association rule mining, Self-Organizing Map（SOM）, audit trail