Title :
An Improved Solution of SOM Network Anomaly Detection Based on T-Distribution
Author :
Weijun, Chen ; Xiao, Li
Author_Institution :
Anyang Normal Univ., Anyang, China
Abstract :
In this paper, a scheme of adaptable distance calculation based on t-distribution is proposed on the basis of analysis of the scheme of SOM network anomaly detection. This method establishes a confidence interval between the test sample and BMU distance using t-distribution. It makes sure that network anomaly occurs when the distance between the test sample and BMU is not within the range of the confidence interval. The improved method is compared with the method of the network anomaly detection based on OC-SVM in order to test its validity. At last, the experimental result shows that this kind of method has characteristics of easy realization, high detection rate and low false alarm rate.
Keywords :
security of data; self-organising feature maps; BMU distance; SOM network anomaly detection; adaptable distance calculation; t-distribution; Data models; Distance measurement; Intrusion detection; Neurons; Support vector machines; Training; Vectors; Self-Organizing Map (SOM); anomaly detection; confidence interval; network security;
Conference_Titel :
Cyber-Enabled Distributed Computing and Knowledge Discovery (CyberC), 2011 International Conference on
Conference_Location :
Beijing
Print_ISBN :
978-1-4577-1827-4
DOI :
10.1109/CyberC.2011.51
