An Audit Method Based on Mathematical Statistics Detection in Database Audit System

According to the method of database audit and the characteristics of current data, this paper design an audit method based on mathematical statistics detection for the audit of database and the control of risk. Through making statistical analysis to user´s behavior, system will generate user´s normal behavior module used in database audit system. This method based on memory mathematical statistics, and on the premise of using hash tables as user´s indexes, it also uses AVL tree to store audit object´s indexes for making it easy to do statistics and configure audit rules. When generating audit rules, this paper uses weighted calculation. By setting appropriate parameters to adjust the role each factor in generating rules, it will generate rules more conforming to the requirements.