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

Table 8

Results of the ablation study on the laboratory dataset.

Method	CBBI w/o FGAN			CBBI
FGAN	✗			✓
Conv-BiLSTM	✓			✓
Device name	Precision	Recall	F-score	Precision	Recall	F-score
360_Camera	0.9504	0.9017	0.9254	0.9697	0.9566	0.9631
Amazon_Echo	0.9023	0.9395	0.9205	0.9435	0.9764	0.9597
D-Link-DSH-C310	0.8882	0.8988	0.8935	1.0000	1.0000	1.0000
Hikvision_DS-IPC-E22H-IW	0.9526	0.9161	0.9340	0.9902	0.9731	0.9816
Hikvision_DS-IPC-S12P-IWT	0.9255	0.9535	0.9393	0.9753	0.9907	0.9829
Ezviz_Camera_CS-C6CN	0.9251	0.9382	0.9316	0.9676	0.9264	0.9465
Ezviz_Door_CS-DB2C	0.9289	0.9596	0.9440	0.9300	0.9856	0.9570
Huawei_Camera_HQ5	0.9966	0.9968	0.9967	0.9988	0.9992	0.9990
Huawei_Speaker	0.9982	0.9936	0.9959	0.9991	0.9978	0.9985
Huawei_Smart_Scale	0.7975	0.8289	0.8129	0.9663	0.9101	0.9373
Imou-Camera_TP1-2525	0.9979	0.9962	0.9971	0.9986	0.9977	0.9982
Imou-Camera_TP7C-E152	0.9894	0.9894	0.9894	0.9977	0.9932	0.9955
Philips_Hue	0.9504	0.9165	0.9332	0.9606	0.9379	0.9491
TP_Link_Camera_IPC42A-4	0.9638	0.9499	0.9568	0.9806	0.9836	0.9821
TP_Link_Camera_IPC43AN-4	0.9581	0.9492	0.9536	0.9881	0.9731	0.9805
TP_Link_Camera_IPC64C-4	0.8520	0.9287	0.8887	0.9360	0.9682	0.9518
TP_Link_WDA6332RE	0.6364	0.4667	0.5385	0.7815	0.8378	0.8087
Xiaomi_Air_Purifier	0.6923	0.3333	0.4500	0.8676	0.8806	0.8741
Xiaomi_Camera_MJSXJ06CM	0.9919	0.9923	0.9921	0.9986	0.9979	0.9983
Xiaomi_Door_LSC_M01	0.9905	0.9881	0.9893	0.9952	0.9967	0.9960
Xiaomi_Hub	0.6857	0.4898	0.5714	0.9975	0.9988	0.9982
Xiaomi_Humidifier	0.8881	0.8975	0.8928	0.9730	0.9689	0.9709
Xiaomi_Soundbox	0.9135	0.9041	0.9088	0.9814	0.9483	0.9646
Macro average performance	0.9033	0.8752	0.8850	0.9651	0.9652	0.9649
Total accuracy	0.9524			0.9726