基于卷积神经网络的免疫网络攻击检测方法  

Immune Network Attack Detection Method Based on Convolutional Neural Network

作　　者：张伟华 王海英 ZHANG Wei-hua;WANG Hai-ying(Zhengzhou Business University,College of Information and Mechanical and Electrical Engineering,Gongyi Henan 451200,China;School of Computer Science and Artificial Intelligence,Wuhan University of Technology,Wuhan Hubei 430000,China)

机构地区：[1]郑州商学院信息与机电工程学院,河南巩义451200 [2]武汉理工大学计算机与人工智能学院,湖北武汉430000

出　　处：《计算机仿真》2025年第2期432-436,共5页Computer Simulation

基　　金：河南省科技厅科技攻关项目(232102220010)。

摘　　要：网络攻击手段不断演化和变异,攻击者会采用各种新的技术和策略来规避检测,加大了攻击检测难度。为此提出基于卷积神经网络的免疫网络攻击检测方法。预处理免疫网络流量数据,采用t-SNE算法降维流量数据,利用Min-Max归一法归一化数据,引入borderline SMOTE算法对数量较少的流量样本过采样,提升免疫网络流量数据质量。结合Dropout正则化方法、随机梯度下降法等改进后卷积神经网络,并将预处理后流量数据用于网络训练,最后利用训练后的卷积神经网络实现免疫网络攻击检测。实验结果表明,所提方法在免疫网络攻击检测中准确率更高、均方误差更小、误检率和漏检率更低、检测效率更高。The methods of network attacks are constantly evolving and mutating,and attackers will adopt various new technologies and strategies to evade detection,increasing the difficulty of attack detection.A convolutional neural network-based immune network attack detection method is proposed for this purpose.Firstly,we pretreated the immune network traffic data,and used t-SNE algorithm to reduce the dimension of traffic data.Secondly,we used Min-Max method to normalize the data.Meanwhile,we introduced borderline SMOTE algorithm to oversample small numbers of traffic samples,thus improving the quality of immune network traffic data.Moreover,we combined Dropout regularization method with random gradient descent method to improve the convolutional neural network.Meanwhile,we used the traffic data after pretreatment for network training.Finally,we used the trained convolutional neural network to realize the immune network attack detection.The experimental results show that the proposed method has higher accuracy,smaller mean squared error,lower false detection rate and missing detection rate,as well as higher detection efficiency in immune network attack detection.

关 键 词：卷积神经网络 免疫网络 攻击检测 数据降维 正则化 

分 类 号：TP393.08[自动化与计算机技术—计算机应用技术]