Network Anomaly Detection Method Based on I-KPCA

DOI

10.2991/cecs-18.2018.4How to use a DOI?

Keywords

I-KPCA, abnormal detection, Nuclear principal component analysis, Guass-Seidel.

Abstract

Network anomaly detection is a hot topic in the field of detection and is of great significance for ensuring the reliable operation of the network. The current research direction is mainly the detection of the host's own operating conditions, and the detection of a single resource, low detection efficiency cannot meet the real-time detection needs and other issues. Based on the theory of Kernel Principal Component Analysis (KPCA), this paper proposes an improved I-KPCA network anomaly detection method, which can integrate multiple data resources for evaluation and greatly reduce the false alarm rate. In order to verify the performance of the detection method, this article focuses on comparative experiments conducted in the Matlab environment. The experimental results show that the network anomaly detection method based on the improved KPCA can not only detect the abnormal situation in real time, but also make the false alarm rate not exceed 0.85% and the detection rate reaches 96%.

Copyright

© 2018, the Authors. Published by Atlantis Press.

Open Access

This is an open access article distributed under the CC BY-NC license (http://creativecommons.org/licenses/by-nc/4.0/).

Download article (PDF)

TY  - CONF
AU  - Jiandong Shang
AU  - Qiang Li
AU  - Runjie Liu
AU  - Yuting Niu
PY  - 2018/07
DA  - 2018/07
TI  - Network Anomaly Detection Method Based on I-KPCA
BT  - Proceedings of the 2018 International Symposium on Communication Engineering & Computer Science (CECS 2018)
PB  - Atlantis Press
SP  - 17
EP  - 21
SN  - 2352-538X
UR  - https://doi.org/10.2991/cecs-18.2018.4
DO  - 10.2991/cecs-18.2018.4
ID  - Shang2018/07
ER  -