摘要
针对入侵检测前必须分析输入数据的特征以及检测中数据维数较高的问题,根据入侵检测的特点,将特征选择问题作为优化问题来考虑,采用量子遗传算法对特征进行选择,充分利用其并行处理及全局搜索能力,提高数据分类质量、降低问题规模、消除冗余属性、加快数据处理速度;在KDD CUP1999数据集上进行实验,结果表明与遗传算法以及粒子群算法相比,该方法可以更有效地精简特征,提高分类质量。
Analysis the characteristics of the input intrusion detection data,and higher dimensions problem of intrusion detection.According to the characteristics of intrusion detection,feature selection will be considered as an optimization problem,using quantum genetic algorithm to feature selection,full use of the quantum genetic algorithm global search and parallel processing capabilities,to eliminate redundant attributes and reduce the scale of the problem and improve the data classification quality,faster data processing speed.Data sets in KDD CUP1999 Experimental results show that genetic algorithms and particle swarm algorithm,this method can more effectively streamline features,improve the quality of classification.
出处
《计算机测量与控制》
CSCD
北大核心
2011年第4期813-815,819,共4页
Computer Measurement &Control
基金
河南省基础与前沿技术研究计划基金项目(082300410390)
关键词
特征选择
入侵检测
量子遗传算法
网络安全
Feature selection
intrusion detection
quantum genetic algorithm
network security
