Divergence based Database Intrusion Detection by user Profile Generation

摘要

For a while Database Intrusion Detection Systems are emerging as a vital requirement. Our objective is to recognize illegitimate outsiders as well as legitimate insiders who are misusing their privileges. Our approach ––– is a novel method to identify malicious transactions by monitoring the individual behavior of a user with the help of feature extraction. A unique User Profile is created for every user by analyzing non-malicious transactions committed by that user. Malicious transactions are detected by utilizing the concept of divergence which examines the deviation between the incoming user and its profile mined earlier. Experiments conducted on a dataset generated using TPC-C benchmark standard has produced promising results.