Author Login Chief Editor Login Reviewer Login Editor Login Remote Office

›› 2005, Vol. 31 ›› Issue (03): 187-189.

• Papers • Previous Articles     Next Articles

Anomaly Detection for Linux Process Behavior Based on Fuzzy ART Neural Network

JI Zhigang, CAI Lidong   

  1. (Department of Computer Science, Jinan University, Guangzhou 510632)
  • Published:2005-02-15

基于Fuzzy ART神经网络的Linux进程行为异常检测

吉治钢,蔡利栋   

  1. 暨南大学计算机科学系,暨南大学计算机科学系 广州 510632

Abstract: In this paper, pattern classification and anomaly detection for Linux process behavior are investigated. The neural network, Fuzzy ART, and its algorithm are introduced, and used to classify patterns and detect anomaly for system calls sequence of Linux process. Primary experiments suggest that this method is feasible. Finally, both merits and demerits of this approach are presented.

摘要: 研究了Linux进程行为的模式提取与异常检测问题。介绍了一种模糊神经网络Fuzzy ART及其实现,利用Fuzzy ART网络对Linux进程的系统调用序列进行模式提取,并据此进行异常检测。实验结果初步表明该方法是可行、有效的。最后说明了该方法的优点和不足。

关键词: FuzzyART, 入侵检测, 异常检测, 神经网络, 系统调用

CLC Number: