摘要
大数据时代的网络安全问题日趋严重,为了提高入侵检测分类的精确性以及节约分类时间资源,构建了基于主成分分析(Principal Component Analysis,PCA)的入侵检测模型。该模型中首先通过主成分分析对入侵信息进行维归约的预处理,降低了入侵信息的维度、压缩了稀疏矩阵,再利用BP算法对预处理后的数据进行分析测试。仿真实验表明,与粗糙集方法相比,PCA方法对入侵信息的维归约效果更佳,它可以在不降低检测率的同时降低入侵信息维度,减少入侵检测的数据量,达到提高入侵检测时间效率的目的,同时节约了网络节点资源。
The network security in the era of large data is becoming more and more serious.In order to improve the accuracy of intrusion detection classification and save the classification time resources,an intrusion detection model based on Principal Component Analysis（PCA） is constructed.In this model,intrusion information dimension reduction preprocessing is carried out by using principal component analysis,the dimension of intrusion information is reduced,and the sparse matrix is compressed;then BP algorithm is used to analyze and test the data after pretreatment.The simulation results show that the PCA method is more effective than the rough set on the intrusion information.PCA method can reduce the intrusion information dimension and the amount of intrusion detection data without reducing the detection rate,improve the efficiency of the intrusion detection time,and save network node resources at the same time.
出处
《河池学院学报》
2017年第5期76-81,共6页
Journal of Hechi University
基金
广西高校中青年教师基础能力提升项目(KY2016YB380)
河池学院智能计算与模式识别重点实验室(校政发[2016]91号)
河池学院重点项目(XJ2017ZD08)
关键词
大数据
入侵检测
主成分分析
降维
big data intrusion detection PC A dimensionality reduction
