Fig 1.
The complete workflow proceeds as follows: 1) Sample edges in the original graph and perform edge interpolation; 2) Using augmented data for contrastive learning; 3) Input the original graph into the GNN model to update the edge attributes; 4) Combining contrastive learning with the edge information learned by the GNN model for traffic detection.
Table 1.
Summary of dataset properties.
Table 2.
Performance of the intrusion detection task for different training set split ratios, expressed as macro-f1 and standard deviation.
Fig 2.
Performance comparison of different methods under various data partitioning schemes.
Table 3.
Performance of the intrusion detection task in the migration scenario, expressed as macro f1 and standard deviation.
Fig 3.
Impact of different sampling sizes in the interpolation method on model performance.
Fig 4.
Impact of different graph construction method on model performance.
Fig 5.
Impact of different graph ratio on model performance.