Research on Database Anomaly Access Detection Based on User Profile Construction

摘要: As a platform for data storage and administration, database contains private large information, which makes it target of malicious personnel attacks. To prevent attacks from outsiders, administrators can limit unauthorized user access through role-based control system, while masquerade insiders are often less noticeable. Therefore, the research on anomaly detection based behavior has important practical application value. In this paper, we proposed system securing database. We took advantage profile construction method to describe query statements without grouping. Then k-means random tree were applied profile. With specified constructed according characteristics submitted by user, is used group users. algorithm train detector. The experimental results show that fast effective detecting behaviors.