Machine learning proposed approach for detecting database intrusions in RBAC enabled databases

Information is valuable asset of any organization which is stored in databases. Data in such databases may contain credit card numbers, social security number or personal medical records etc. Failing to protect these databases from intrusions will result in loss of customer's confidence and might even result in lawsuits. Traditional database security mechanism does not design to detect anomalous behavior of database users. There are number of approaches to detect intrusions in network. But they cannot detect intrusions in database. There have been very few ID mechanisms specifically tailored to database systems. We propose transaction level approach to detect malicious behavior in database systems enabled with Role Based Access Control (RBAC) mechanism.