Real Network Traffic Collection and Deep Learning for Mobile App Identification

Table 2

Accuracy and macroaverage F-measure for the baseline and DL models on two datasets, respectively.

Model		Dataset-all		Dataset-TLS
		Accuracy (%)	Macro-F1 (%)	Accuracy (%)	Macro-F1 (%)
Baseline	RF	61.32	50.30	62.91	50.07
	MLP	81.24	70.64	83.57	65.79
DL	SDAE	58.73	39.86	60.37	41.45
	1D CNN	91.80	90.10	93.14	84.25
	LSTM	85.09	80.77	89.36	80.51