To address the issues of limited feature extraction,strong data dependency,and poor generalization ability in existing network intrusion detection methods,this paper proposes an intrusion detection method based on self-supervised graph neural net-works and hybrid neural networks.Firstly,through a self-supervised learning strategy,a graph convolutional network is employed to extract structural features from network traffic data,enhancing the model's ability to learn features from unlabeled data.This re-duces dependence on labeled data and improves generalization ability.Secondly,a convolutional neural network is used to extract spatial features from the time series of network traffic,and a long short-term memory network is employed to model temporal de-pendencies,enabling multi-view feature extraction and improving detection comprehensiveness.Finally,a feature fusion strategy is designed to enrich the model's feature representation and enhance its robustness.Experimental results on public datasets dem-onstrate that the proposed method achieves higher accuracy and F1 score.
维普
万方数据