Differentially Private Tabular Synthetic Data Generation on SCM simulation RFF prior (test)
81.12AccuracyUB
Evaluation Results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| UBPrior=RFF, epsilon=1.02026.06 | 81.12 | 81.12 | 0.113 | 0.152 | 0.191 | 98.12 | 97.55 | |
| Tab-PEPrior=RFF, epsilon=1.02026.06 | 64.1 | 69.19 | 0.12 | 0.16 | 0.198 | 98.57 | 81.27 | |
| PrivMRFPrior=RFF, epsilon=1.02026.06 | 60.11 | 63.68 | 0.122 | 0.162 | 0.202 | 93.06 | 98.55 | |
| AIMPrior=RFF, epsilon=1.02026.06 | 59.6 | 62.76 | 0.125 | 0.165 | 0.204 | 93.41 | 98.5 | |
| RAP++Prior=RFF, epsilon=1.02026.06 | 59 | 62.13 | 0.178 | 0.221 | 0.263 | 0.27 | 25.84 | |
| PrivGSDPrior=RFF, epsilon=1.02026.06 | 57.08 | 59.7 | 0.137 | 0.179 | 0.22 | 85.13 | 90.8 | |
| GEMPrior=RFF, epsilon=1.02026.06 | 53.76 | 55.53 | 0.187 | 0.243 | 0.298 | 57.8 | 98.51 | |
| PrivSynPrior=RFF, epsilon=1.02026.06 | 50.96 | 50.56 | 0.173 | 0.227 | 0.281 | 66.83 | 97.77 |