An adaptive KNN algorithm for anomaly intrusion detection

Abstract

Intrusion detection is the process of monitoring the activity in the target system and analyzing it for intrusive activity. The purpose of the present paper is to create a methodology for the anomaly-based intrusion detection, which is grounded on the kd trees for the description of the system activity and the k-Nearest Neighbor algorithm during the intrusion detection phase, as well as the evaluations of the results.