Table 2.
10-fold cross-validated SAYL performance. AUC is Area Under the Curve. Rule number averaged over the 10 folds of theories. For comparison, we include results of Differential Prediction Search (DPS) and Model Filtering (MF) methods [20].
| Algorithm | Uplift AUC |
Lift(older) AUC |
Lift(younger) AUC |
Rules Avg # |
DPS p-value |
|
| SAYL | 58.10 | 97.24 | 39.15 | 9.3 | 0.002 | * |
| DPS | 27.83 | 101.01 | 73.17 | 37.1 | - | |
| MF | 20.90 | 100.89 | 80.99 | 19.9 | 0.0039 | * |
| Baseline | 11.00 | 66.00 | 55.00 | - | 0.0020 | * |
We compute the p-value comparing each method to DPS, * indicating significance.