Title :
Reason for Hierarchical Self Organized Map-Based Intrusion Detection System Incapable of Increasing Detection Rate
Author_Institution :
Coll. of Comput. Sci. & Eng., Chongqing Univ., Chongqing, China
Abstract :
When the single-layered self organized map (SOM) is used for anomaly detection, ldquooverlaprdquo exists for five types of samples. The hierarchical SOM system, which increases classification granularity, can not effectively enhance the detection rate. Through analysis to clustering details in the hierarchical SOM system, the reason for a low detection rate of attack was found, and improvement were proposed based on the reason.
Keywords :
security of data; self-organising feature maps; anomaly detection; classification granularity; hierarchical self organized map-based intrusion detection system; single-layered self organized map; Computer science; Computer vision; Educational institutions; Electronic commerce; Intrusion detection; Neural networks; Neurons; Spatial databases; Usability; Watches; Hierarchical SOM; Intrusion Detection; Remote to Local attack;
Conference_Titel :
Information Engineering and Electronic Commerce, 2009. IEEC '09. International Symposium on
Conference_Location :
Ternopil
Print_ISBN :
978-0-7695-3686-6
DOI :
10.1109/IEEC.2009.36
