OpenCBD: A Network-Encrypted Unknown Traffic Identification Scheme Based on Open-Set Recognition

Table 7

When the number of encoders in the ensemble model of OpenCBD is 4, 8, and 12, accuracy, precision, recall, and F1-score of the 9-class classification tasks (identifying 8 known classes and 1 unknown class) vary as the number of known classes is randomly selected in the first training stage.

	Accuracy			Precision			Recall			F1-score
4	8	12	4	8	12	4	8	12	4	8	12
3	0.5315	0.5715	0.5838	0.6462	0.6717	0.6893	0.4593	0.5953	0.6051	0.5019	0.6144	0.6258
4	0.5961	0.653	0.656	0.6241	0.7137	0.723	0.5988	0.6784	0.6813	0.5994	0.6863	0.6929
5	0.6084	0.7007	0.6976	0.6991	0.7425	0.7355	0.6104	0.7446	0.7197	0.6343	0.7385	0.7227
6	0.6646	0.713	0.703	0.7419	0.7704	0.772	0.6613	0.7242	0.7186	0.6845	0.7395	0.7339
7	0.7046	0.7223	—	0.7639	0.7616	—	0.7315	0.7597	—	0.7404	0.7533	—