A novel method for intrusion detection in relational databases

The importance of data security and confidentiality increases day by day, since for most companies and organizations data remains as the most important asset. Standard database security measures like access control mechanisms, authentication and encryption technologies are of little help when it comes to preventing data theft from insiders. By incorporating intrusion detection mechanisms, we can improve the security features of a Database Management System (DBMS). In this paper we propose a novel method for detecting intrusions in databases using data mining techniques like clustering and classification. Experiments show that our method outperforms other methods with higher accuracy and reduced false alarm rate.