基于DL和TSVM的入侵检测方法研究  

INTRUSION DETECTION METHOD BASED ON DL AND TSVM

作　　者：魏明军 彭宁 Wei Mingjun;Peng Ning(College of Information Engineering,North China University of Science and Technology,Tangshan 063210,Hebei,China)

机构地区：[1]华北理工大学信息工程学院,河北唐山063210

出　　处：《计算机应用与软件》2020年第12期328-333,共6页Computer Applications and Software

摘　　要：为解决网络环境下大量高维数据给入侵检测造成的数据特征提取不当、检测速度慢、检测率低的问题,提出一种基于深度置信网络(Deep Belief Network,DBN)和孪生支持向量机(Twin Support Vector Machine,TSVM)的入侵检测模型(DBN-TSVM-5)。利用五层受限玻尔兹曼机的DBN对归一化后的标准数据集进行特征降维,以获得入侵检测数据的最优低维表示,构造多分类TSVM-5分类器,对降维后的数据进行识别。经过KDDCUP99数据集的仿真实验,结果表明,该模型是一种有效的入侵检测模型。An intrusion detection model(DBN-TSVM-5)based on deep belief network(DBN)and twin support vector machine(TSVM)is proposed to solve the problems of improper data feature extraction,slow detection speed and low detection rate caused by a large amount of high-dimensional data in network environment.In this method,DBN of five-layer restricted boltzmann machine was used to reduce the characteristic dimension of the normalized standard data set to obtain the optimal low-dimensional representation of intrusion detection data.Multi-classification TSVM-5 classifier was constructed to identify the data after dimension reduction.The simulation results on KDDCUP99 data set show that the model is an effective intrusion detection model.

关 键 词：深度学习 入侵检测 深度置信网络 孪生支持向量机 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]