Differentially private sum estimation on Pages data
4,324.38MAEEMQ
Evaluation results
| Method | Links | ||
|---|---|---|---|
| EMQPrivacy (epsilon)=1, q (quantile)=best of {0.95, 0.96, 0.97, 0.98, 0.99}2023.05 | 4,324.38 | 2,343.25 | |
| ATPrivacy (epsilon)=1, q (quantile)=best of {0.95, 0.96, 0.97, 0.98, 0.99}2023.05 | 4,377.45 | 2,340.93 | |
| UQEPrivacy (epsilon)=1, q (quantile)=0.992023.05 | 4,385.23 | 2,077.16 | |
| UQEPrivacy (epsilon)=0.5, q (quantile)=0.992023.05 | 7,102.34 | 3,093.13 | |
| ATPrivacy (epsilon)=0.5, q (quantile)=best of {0.95, 0.96, 0.97, 0.98, 0.99}2023.05 | 7,491.97 | 4,367.31 | |
| EMQPrivacy (epsilon)=0.5, q (quantile)=best of {0.95, 0.96, 0.97, 0.98, 0.99}2023.05 | 7,601.47 | 3,963.52 | |
| UQEPrivacy (epsilon)=0.1, q (quantile)=0.992023.05 | 21,916.37 | 6,423.75 | |
| ATPrivacy (epsilon)=0.1, q (quantile)=best of {0.95, 0.96, 0.97, 0.98, 0.99}2023.05 | 30,830.49 | 20,422.05 | |
| EMQPrivacy (epsilon)=0.1, q (quantile)=best of {0.95, 0.96, 0.97, 0.98, 0.99}2023.05 | 46,552.09 | 27,944.18 |