Skip to main content
. 2022 Aug 3;5(1):26. doi: 10.1186/s42400-022-00129-6

Fig. 13.

Fig. 13

The MSE averaged over 100 runs for synthetic non-uniform data with N = 500,000, D = 20 privatized using different randomization techniques and fake data under a high privacy regime, experimental setup labeled ES_Syn_6 in Table 2; b general privacy regime, experimental setup labeled ES_Syn_12 in Table 2