Title :
SVC-Based Multivariate Control Charts for Automatic Anomaly Detection in Computer Networks
Author :
Zhisheng Zhang ; Xuejun Zhu
Author_Institution :
Southeast Univ., Nanjing
Abstract :
The design of multivariate control charts for automatic anomaly detection in computer networks is a challenging research issue due to the complexity of the data structure of the network operational data. In general, the design of statistical multivariate control charts is limited to a Gaussian distribution assumption or a pre-known probability distribution model, which is hardly applicable to the computer operation data. The paper is motivated by this timely need to develop SVC (support vector clustering) based multivariate control charts, which do not require the data to have a pre-known probability distribution model. The proposed method is validated through the simulations by comparing with the popularly used statistical T2 multivariate control charts. The effectiveness of the method is also demonstrated through automatic anomaly detection of typical computer intrusions.
Keywords :
computer networks; control charts; data structures; optimisation; pattern clustering; statistical analysis; support vector machines; telecommunication computing; telecommunication security; unsupervised learning; SVC-based multivariate control charts; automatic anomaly detection; computer networks; data structure; optimization problem; support vector clustering; unsupervised kernel based control charts; Computer hacking; Computer industry; Computer networks; Control charts; Data engineering; Design engineering; Industrial control; Mechanical engineering; Probability distribution; Static VAr compensators; control chart; intrusion detection; machine learning;
Conference_Titel :
Autonomic and Autonomous Systems, 2007. ICAS07. Third International Conference on
Conference_Location :
Athens
Print_ISBN :
978-0-7695-2859-7
Electronic_ISBN :
978-0-7695-2859-7
DOI :
10.1109/CONIELECOMP.2007.99
