Semisupervised Graph Neural Networks for Traffic Classification in Edge Networks

Table 1

Related works based on deep learning models (single-modal architectures).

Research	Input data	TO	DL classifiers	O
Wang et al. [10], proc. IEEE ISI	(1) PCAP trace	F/B	1D-CNN	Y
	(2) First 784 bytes of L4 payload
Lopez-Martin et al. [12], IEEE access	6 fields extracted per packets and first 20 packets per biflow	B	2D-CNN; +LSTM	N
Li et al. [13], IWQoS	IP datagram	D	RNN	N
Liu et al. [14], proc. IEEE INFOCOM	IP packet lengths	B	BiGRU	N
Lotfollahi et al. [11], soft computing	First 1500 bytes of L2 payload	P	SAE/1D-CNN	Y