Cohort Analytics on Cohort Analytics 1.0 (Evaluation)
0.91P-VaR 0.95Hybrid (k-anonymity + DP)
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| Hybrid (k-anonymity + DP)k=200, epsilon=0.1, Guarantee=Both, Cost=O(n log n), Interp.=High2026.01 | 0.91 | 0.1 | 0.79 | |
| Hybrid (k-anonymity + DP)k=200, epsilon=0.3, Guarantee=Both, Cost=O(n log n), Interp.=High2026.01 | 1.26 | 0.3 | 0.91 | |
| Differential Privacy (DP)epsilon=0.1, Guarantee=ε-DP, Cost=O(n), Interp.=Medium2026.01 | 1.34 | 0.1 | 0.72 | |
| Rényi DPalpha=32, Guarantee=(α, ε)-RDP, Cost=O(n log n), Interp.=Medium2026.01 | 1.58 | 0.3 | 0.89 | |
| Hybrid (k-anonymity + DP)k=100, epsilon=0.3, Guarantee=Both, Cost=O(n log n), Interp.=High2026.01 | 1.63 | 0.3 | 0.89 | |
| Hybrid (k-anonymity + DP) + Synthetic Fallbackk=100, epsilon=0.3, synthetic_fallback=true, Guarantee=Both + FB, Cost=O(n log n), Interp.=High2026.01 | 1.63 | 0.3 | 0.89 | |
| Differential Privacy (DP)epsilon=0.5, Guarantee=ε-DP, Cost=O(n), Interp.=Medium2026.01 | 2.18 | 0.5 | 0.91 | |
| Hybrid (k-anonymity + DP)k=50, epsilon=0.5, Guarantee=Both, Cost=O(n log n), Interp.=High2026.01 | 2.41 | 0.5 | 0.91 | |
| Differential Privacy (DP)epsilon=1, Guarantee=ε-DP, Cost=O(n), Interp.=Medium2026.01 | 3.12 | 1 | 0.94 | |
| k-anonymityk=100, Guarantee=Indisting., Cost=O(n log n), Interp.=High2026.01 | 3.87 | 6.64 | 0.98 | |
| Local DPepsilon=4, Guarantee=Local ε-DP, Cost=O(n), Interp.=Low2026.01 | 3.87 | 4 | 0.82 | |
| k-anonymityk=50, Guarantee=Indisting., Cost=O(n log n), Interp.=High2026.01 | 4.23 | 5.64 | 0.98 | |
| No ProtectionGuarantee=None, Cost=O(n), Interp.=High2026.01 | 8.45 | — | 1 |