Real-time trajectory privacy protection based on improved differential privacy method and deep learning model

Journal of Cloud Computing

Advances, Systems and Applications

Table 5 Privacy protection average performance of different query ranges

Dataset	Query Range	Query Error			Jensen–Shannon Divergence			Kendall Coefficient
Dataset	Query Range	UG	AG	Top-Down	UG	AG	Top-Down	UG	AG	Top-Down
Yonsei	4*4	0.00016	0.00011	0.0001	0	0.0314	0.0787	0	0	0
	8*8	0.00074	0.00032	0.00019	4.3482	1.8328	0.3173	0	0	0
	16*16	0.00086	0.00063	0.00058	5.3425	2.6006	1.0003	0	0	0
	32*32	0.0011	0.00067	0.00034	4.9071	1.7646	0.8035	0.0014	0.0018	0.0021
	64*64	0.0008	0.00064	0.00042	4.3649	1.1301	0.003	0.0051	0.0057	0.0064
Citi Bike	4*4	0.00065	0.00059	0.0005	0.3271	0.7985	1.5057	0	0	0
	8*8	0.0008	0.00073	0.00056	4.6952	4.1017	3.2115	0	0	0
	16*16	0.0011	0.00096	0.0009	4.8393	4.5999	4.2408	0	0	0
	32*32	0.00101	0.00106	0.0011	4.6977	4.5786	4.4001	0	0	0
	64*64	0.00106	0.00109	0.0009	0.1748	0.1057	0.0022	0.0212	0.0353	0.0447