基于卷积神经网络的网络入侵检测技术  

Network Intrusion Detection Technology Based on Convolutional Neural Network

作　　者：黄志敏 彭世强 HUANG Zhimin;PENG Shiqiang(Guangzhou Chinagdn Security Technology Co.,Ltd.,Guangzhou 510665,China)

机构地区：[1]广州竞远安全技术股份有限公司,广东广州510665

出　　处：《电子质量》2024年第1期10-15,共6页Electronics Quality

摘　　要：随着互联网的发展,网络安全问题是互联网发展所面临的一个严峻挑战,网络入侵检测技术成为其中需要重点关注的问题。特别是随着攻击手段的进一步多样化和数据维度的不断增加,传统的机器学习算法已不能满足目前网络入侵检测系统的要求。卷积神经网络(CNN)具有强大的特征提取能力和数据分析能力,可以提高网络入侵检测的准确性和时效性。因此将CNN应用到网络入侵检测技术中,并通过交叉熵损失函数达到提升检测准确率的目的。首先,对公开数据集进行预处理;然后,构建CNN模型获取分类预测结果;最后,计算模型评价指标,并不断调整CNN模型,直到模型评价指标达到期望值。With the development of the Internet,cybersecurity is a serious challenge for the development of the Internet,and network intrusion detection technology has become one of the issues that need to be focused on.Especially with the further diversification of attack methods and the continuous increase of data dimension,the traditional machine learning algorithms can no longer meet the requirements of current network intrusion detection system.Convolutional neural networks(CNN)have powerful feature extraction capabilities and data analysis capabilities,which can improve the accuracy and timeliness of network intrusion detection.Therefore,CNN is applied to network intrusion detection technology,and the detection accuracy is improved through the cross entropy loss function.Firstly,the public dataset is preprocessed.Then,the CNN model is constructed to obtain the classification prediction results.Finally,the model evaluation index is calculated,and the CNN model is continuously adjusted until the expected value of the model evaluation index is reached.

关 键 词：网络入侵检测 卷积神经网络 深度学习 机器学习 

分 类 号：TP393.08[自动化与计算机技术—计算机应用技术]