Research Article

Identifying IoT Devices Based on Spatial and Temporal Features from Network Traffic

Table 7

Results of the ablation study on the UNSW dataset.

MethodCBBI w/o FGANCBBI

FGAN
Conv-BiLSTM
Device namePrecisionRecallF-scorePrecisionRecallF-score
Amazon_Echo0.98780.99320.99050.99830.99880.9985
Belkin_Wemo_Switch0.99320.99290.99310.99820.99880.9985
HP_Printer0.90560.96460.93420.99080.99810.9944
Insteon_Camera0.99310.99810.99560.99940.99940.9994
Light_Bulbs_LiFX_Smart_Bulb0.98250.98250.98250.99790.99650.9972
Netatmo_Weather_Station0.98650.99630.99140.99881.00000.9994
Netatmo_Welcome0.91810.95330.93540.99370.99560.9947
PIX_STAR_Photo_Frame0.99260.98870.99070.99900.99510.9971
Samsung_SmartCam0.99780.99220.99500.99920.99890.9991
Smart_Things0.99560.98880.99220.99850.99800.9983
TP_Link_Day_Night_Cloud_Camera0.98980.95630.97280.98920.99890.9940
TP_Link_Smart_Plug0.86380.86750.86570.98040.96150.9709
Withings_Aura_Smart_Sleep_Sensor0.99290.97930.98610.99890.99850.9987
Withings_Smart_Baby_Monitor0.98230.99610.98910.99970.99970.9997
iHome1.00000.68330.81191.00001.00001.0000
Nest_Dropcam0.31930.43710.39801.00000.96000.9796
NEST_Procet_Smoke_Alarm1.00000.73170.84511.00001.00001.0000
Triby_Speaker0.88410.45520.60100.98170.94690.9640
Macro average performance0.93250.88650.90390.99580.99140.9935
Total accuracy0.98650.9983