Advances, Systems and Applications
Table 1 Related work classification table
Method | Data | Algorithm | Technical characteristics |
|---|---|---|---|
HS-Trees [25] | No reliance | ML | The detection of anomaly is quickly achieved by decision tree that does not require changes in tree structure. |
RShash [26] | ML | Using random hashing to detect subspace anomaly | |
LSTM [27] | CNN | Using LSTM AutoEncoder pairs to extract Feature, and use SVM to complete the binary classification of input. | |
ICNAD [28] | GNN | Network anomaly detection is achieved using GNN that fuse the nodes’ attributes and neighboring nodes’. | |
GDN [29] | GNN | Improving the accuracy of anomaly detection by using GAT with graph neural network. | |
Snappy [21] | Network Telemetry | Statistical Analysis | A Microburst detection method implemented inside programmable switches. |
BurstRadar [30] | Statistical Analysis | A Microburst monitoring algorithm inside programmable switches for real-time detection. | |
INT-DETECT [31] | ML | Grey fault detection and localization method based on INT. | |
PacketScope [32] | Statistical Analysis | Determine whether abnormal packet loss events occur within network based on INT data. | |
INT-detector [33] | GAAL | Fast network anomaly detection based on INT and GAAL. | |
LossSight [34] | GAN | Packet loss complementary based on INT and GAN. | |
ODS [35] | Clustering | Detection of BGP anomalies based on clustering algorithm. |