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

Table 8

When the number of encoders in the ensemble model of OpenCBD is 4, 8, and 12, accuracy, precision, recall, and F1-score of the 2-class classification tasks (identifying known and unknown classes) vary with increasing percentage of known classes randomly selected.

	Accuracy			Precision			Recall			F1-score
4	8	12	4	8	12	4	8	12	4	8	12
60%	0.7046	0.72	0.7184	0.753	0.7678	0.7726	0.7737	0.7812	0.7687	0.7632	0.7744	0.7706
70%	0.7038	0.7423	0.773	0.767	0.8015	0.8371	0.745	0.7725	0.7837	0.7558	0.7867	0.8095
80%	0.7046	0.7323	0.7423	0.7619	0.786	0.8047	0.7563	0.7762	0.7675	0.759	0.7811	0.7856
90%	0.7384	0.7315	0.7538	0.7875	0.7909	0.8269	0.7875	0.7662	0.7587	0.7875	0.7784	0.7913