摘要
如何提取数据的有效特征是入侵检测过程中至关重要的方法。本文针对常用的特征选择算法无法很好地对入侵检测数据特征进行降维以及检测时间长等不足,提出了一种基于主成分分析法的入侵检测特征选择方法。首先利用PCA方法对原始的数据集合进行正交变换,然后根据变换所得的矩阵,分析每个原始特征在正交变换时的贡献度,最后依据贡献度的大小对原始特征属性进行选择。通过仿真实验,表明通过PC A方法所选择出的特征变量涵括了入侵检测数据特征的主要信息,属于重要特征。在保证检测率的基础上减少了检测时间,提高了检测效率。
In the process of intrusion detection,how to select the effective data feature is an important link of the process of detection.Commonly used feature selection methods are not able to reduce dimension so good that detection time too long. According to this,This paper presents a method to intrusion detection feature selection based on principal component analysis. It carries on the PCA transformation to the original data. Then analysising the weight of each component by the transform matrix. Finally,select the most representative feature from the original by the transform weight. The experimental results show that the features selected by the method presented in this paper are enough representative. At the same time the detection time is reduced and the detection efficiency is improved on the basis of ensuring the detection rate.
作者
刘婷
刘晓洁
岳未然
LIU Ting LIU Xiaojie YUE Weiran(Computer college, Sichuan University, Chengdu, 610041, China)
出处
《网络新媒体技术》
2017年第2期28-32,共5页
Network New Media Technology
基金
国家重点研发计划(2016yfb00604
2016yfb00605)
国家自然科学基金项目(61572334)
关键词
主成分分析
入侵检测
特征选择
principal component analysis
intrusion detection
feature selection
作者简介
刘婷,1991年生,女,硕士研究生,主要研究方向是网络与信息安全。
岳未然,1992年生,男,硕士研究生,主要研究方向是网络与信息安全。
通信作者:刘晓洁。