Computer Engineering and Applications ›› 2013, Vol. 49 ›› Issue (7): 89-92.

Previous Articles     Next Articles

Intrusion feature selection algorithm based on Particle Swarm Optimization

WU Qingtao, CAO Jibang, ZHENG Ruijuan, ZHANG Juwei   

  1. Electronic and Information Engineering College, Henan University of Science and Technology, Luoyang, Henan 471003, China
  • Online:2013-04-01 Published:2013-04-15

基于粒子群优化的入侵特征选择算法

吴庆涛,曹继邦,郑瑞娟,张聚伟   

  1. 河南科技大学 电子信息工程学院,河南 洛阳 471003

Abstract: Intrusion feature selection can improve the correctness and detection rate of the intrusion detection system effectively. A intrusion feature subset selection algorithm based on Particle Swarm Optimization(PSO) is proposed. Depending on the analyses of correlation between all features of network intrusion data, the PSO algorithm is used to search and choose effective feature subset independently to reduce data dimension in the feature space. The experimental results show that the algorithm can effectively remove redundant features, reduce the time of feature selection, ensure detection accuracy and improve detecting speed.

Key words: intrusion detection, Particle Swarm Optimization(PSO), intrusion feature selection, optimization searching, feature relevance

摘要: 针对高维数入侵检测数据集中信息冗余导致入侵检测算法处理速度慢的问题,提出了一种基于粒子群优化的入侵特征选择算法,通过分析网络入侵数据特征之间的相关性,可使粒子群优化算法在所有特征空间中优化搜索,自主选择有效特征子集,降低数据维度。实验结果表明该算法能够有效去除冗余特征,减少特征选择时间,在保证检测准确率的前提下,有效地提高了系统的检测速度。

关键词: 入侵检测, 粒子群优化, 特征选择, 优化搜索, 特征关联性