| 基于ResNet BiGRU的入侵检测研究 |
 点此下载全文 |
| 引用本文：刘东, 柳毅.基于ResNet BiGRU的入侵检测研究[J].计算技术与自动化,2025,(2):9-14 |
| 摘要点击次数: 1027 |
| 全文下载次数: 769 |
|
|
| 中文摘要:随着网络流量的爆炸性增长和入侵行为的不断多样化，传统的入侵检测方法在面对海量的网络数据时，存在特征提取不充分、模型存在过拟合、检测分类的准确率不足等问题，提出了一种基于ResNet BiGRU网络的入侵检测算法。该算法首先使用SMOTE过采样重构训练数据集以缓解正负类样本不平衡问题，提出了一种网络并行的结构充分提取特征，将卷积神经网络和双向门控单元同时提取空间尺度特征和时间序列特征，为了将空间尺度特征和时间尺度特征更好地聚合表达，设计了一个Transformer混合融合块，最后利用Softmax函数进行分类。为验证模型的有效性，在UNSW NB15数据集进行实验，对比结果表明该模型在二分类和多分类任务上检测准确率分别达到94.1%和82.6%，结果表明提出的算法具有较高的性能和有效性。 |
| 中文关键词:网络安全 双向门控循环单元 残差网络 混合融合块 入侵检测 |
| |
| Research on Intrusion Detection Based on ResNet BiGRU |
|
|
| Abstract:With the explosive growth of network traffic and the continuous diversification of intrusion behaviors, traditional intrusion detection methods suffer from insufficient feature extraction, overfitting of models, and insufficient accuracy of detection and classification in the face of massive network data.An intrusion detection algorithm based on ResNet BiGRU network is proposed. The algorithm first uses SMOTE oversampling to reconstruct the training dataset to alleviate the positive and negative class sample imbalance problem, proposes a network parallel structure to extract features adequately, combines the convolutional neural network and the bidirectional gating unit to extract both spatial scale features and time series features, and in order to express the spatial scale features and the temporal scaled features in a better way by aggregating them, designs a Transformer hybrid fusion block, and finally the Softmax function is utilized for classification. In order to verify the effectiveness of the model, experiments are carried out on the UNSW NB15 dataset, and the comparison results show that the model achieves detection accuracy of 94.1% and 82.6% on binary and multi classification tasks, respectively, and the results indicate that the proposed algorithm has high performance and effectiveness. |
| keywords:networks security bi directional gated recurrent units residual networks hybrid fusion block intrusion detection |
| 查看全文 查看/发表评论 下载pdf阅读器 |
|
|
|
位访问者,今日访问
次