Detection and Analysis of Intrusion Characteristic Based on BP Neural Network

Abstract:

A Detection and Analysis algorithm base on BP neural network was proposed to solve the problem of low intercept rate, which was exist in traditional algorithm. The traditional algorithm cannot detect the intrusion effectively because the intruder was always combined with other computer virus. The BP neural network can transfer the input to output with nonlinear method, the characteristic was extracted and compare, then the intruder can be detected and intercepted with higher probability. Compared with traditional algorithm, the new algorithm can intercept the intruder with 22% higher rate. Result shows good performance of the algorithm in the intrusion signal detection and excellent perspective in practice.

• Periodical: Advanced Materials Research (Volumes 846-847)
• Pages: 1720-1723
• DOI: https://doi.org/10.4028/www.scientific.net/AMR.846-847.1720
• Online since: November 2013
• Authors: Hong Liang Guo, Shao Ying Kong*
• Keywords: BP Neural Network (BPNN), Characteristic Detection, Computer Network Intrusion

References

[1] Yan Qianhe, Yan Zhenping. Optimization neural network intrusion detection system based on genetic algorithm [J]. Computer Simulation, 2011, 28(4): 141-144.

[2] Ding Ling, Zhao Xiaogang. Intrusion detection base on improved BP neural network [J]. Micro computer information, 2012, 28(3): 131-134.

[3] Wang Lihui. A linkage firewall network intrusion detection system [J]. Application research of computers. 2006, 3: 95-97.

[4] China internet network information center, China Internet Development Statistics Report [R], 2008. 6.