Combined data mining approach for intrusion detection

Abstract

This paper presents the results of the project MIAU, a data mining approach for intrusion detection alert correlation. MIAU combines different data mining techniques in order to properly solve some existing problems in the management and analysis of alerts generated by actual intrusion detection systems. Some of these data mining methods and their application to MIAU are introduced in this paper. Experiments have been carried out with the purpose of demonstrating the validity of the proposed model and some conclusions about them are extracted. Finally, some possible improvements for the system and further work are exposed.