Share Email Print
cover

Proceedings Paper

A novel unsupervised anomaly detection based on robust principal component classifier
Author(s): Wenbin Qiu; Yu Wu; Guoyin Wang; Simon X. Yang; Jie Bai; Jieying Li
Format Member Price Non-Member Price
PDF $14.40 $18.00

Paper Abstract

Intrusion Detection Systems (IDSs) need a mass of labeled data in the process of training, which hampers the application and popularity of traditional IDSs. Classical principal component analysis is highly sensitive to outliers in training data, and leads to poor classification accuracy. This paper proposes a novel scheme based on robust principal component classifier, which obtains principal components that are not influenced much by outliers. An anomaly detection model is constructed from the distances in the principal component space and the reconstruction error of training data. The experiments show that this proposed approach can detect unknown intrusions effectively, and has a good performance in detection rate and false positive rate especially.

Paper Details

Date Published: 18 April 2006
PDF: 9 pages
Proc. SPIE 6241, Data Mining, Intrusion Detection, Information Assurance, and Data Networks Security 2006, 62410T (18 April 2006); doi: 10.1117/12.664383
Show Author Affiliations
Wenbin Qiu, Chongqing Univ. of Posts and Telecommunications (China)
Yu Wu, Chongqing Univ. of Posts and Telecommunications (China)
Guoyin Wang, Chongqing Univ. of Posts and Telecommunications (China)
Simon X. Yang, Univ. of Guelph (Canada)
Jie Bai, Chongqing Univ. of Posts and Telecommunications (China)
Jieying Li, Chongqing Univ. of Posts and Telecommunications (China)


Published in SPIE Proceedings Vol. 6241:
Data Mining, Intrusion Detection, Information Assurance, and Data Networks Security 2006
Belur V. Dasarathy, Editor(s)

© SPIE. Terms of Use
Back to Top