火力与指挥控制2025，Vol.50Issue(6)：21-27,7.DOI:10.3969/j.issn.1002-0640.2025.06.003
融合CNN-GRU和Transformer的网络入侵检测方法
Networkintrusion Detection Method Based on Fusion of CNN-GRU and Transformer
摘要
Abstract
With the rapid development of network technology and its widely application in the military field,and IDS is very important to the security of the system.To address the issue of class imbalance in traditional intrusion detection datasets,a network intrusion detection method called CNN-GRU transformer(CGT)is proposed,which combines convolutional gated recurrent units(CNN-GRU)and a neural network model based on self-attention mechanism(Transformer).This method optimizes intrusion detection technology by addressing the drawbacks of bidirectional long short-term memory(LSTM)networks,which only consider temporal features while neglecting spatial features and have a large number of parameters.Additionally,the dataset is balanced using the NBW model(neighbourhood-cleaning-rule borderline-SMOTE WGAN),which combines over-and under-sampling with wasserstein generative adversarial networks.The experimental results demonstrate that the proposed method exhibits promising performance on the NSL-KDD dataset,effectively enhancing intrusion detection performances.关键词
入侵检测/卷积门控循环单元/数据平衡处理/领域清理规则/神经网络Key words
intrusion detection/convolutional gated recurrent unit(ConvGRU)/data balancing handling/domain cleaning rule/neural network分类
信息技术与安全科学引用本文复制引用
黄迎春,邢秀祺..融合CNN-GRU和Transformer的网络入侵检测方法[J].火力与指挥控制,2025,50(6):21-27,7.基金项目
国家自然科学基金资助项目(61971291) （61971291）
