Table 6. MIA result summary of trained attack on Breast Cancer dataset for heavy noise level.

Heavy noise ( $\epsilon =0.5,\delta ={10}^{-3}$) Attack type = trained attack
Slice feature/	Attacker	AUC	Precision	Recall	F1 score
Slice value	advantage
Entire dataset	0.30	0.42	0.55	0.47	0.51
Class 0	0.24	0.39	0.57	0.47	0.51
Class 1	0.19	0.44	0.52	0.47	0.50
Correctly classified	0.12	0.46	0.49	0.45	0.47
True
Correctly classified	0.36	0.54	0.49	0.56	0.52
False