An Intrusion Detection Syst System Based on Hierarchicalem Self-Organization

摘要

An intrusion detection system (IDS) mon monitors the IP packets flowing over the net- itors networkto capture intrusions or anomalies. One of the techniques used for anomaly detection iswork building statistical models using metrics derived from observation of the user's actions. A neur- neuralnetwork model based on self organization is proposed for detecting intrusions. The self- al selforganizingmap (SOM) has shown to be successful for the analysis of high-dimensional inputorganizing data as in data mining applications such as network security. The proposed growing hierarchic- hierarchicalSOM (GHSOM) addresses the limitations of the SO al SOM related to the static architecture of thisM model. The GHSOM is an artificial neural network model with hierarchical architecture com- composedof independent growing SOMs. Randomly selected subsets that contain both attacks andposed normal records from the KDD Cup 1999 benchmark are used for training the proposedGHSOM.