Deep-Feature-Based Autoencoder Network for Few-Shot Malicious Traffic Detection

Table 1

Experiments results of dataset CIC-IDS2017.

Methods	With unknown categories			Known categories
	Recall	Precision	Macro-f1	Recall	Precision	Macro-f1
Similarity measure	0.5233	0.5134	0.5103	0.4134	0.4230	0.4071
K-means cluster	0.6428	0.5554	0.3228	0.6163	0.6962	0.4721
AE	0.6226	0.5929	0.6042	0.8510	0.8148	0.8196
CNN with AE	0.6718	0.6399	0.6539	0.8626	0.8262	0.8323
DFAE	0.8593	0.8738	0.8579	0.9519	0.9508	0.9513