Table 1.
Winning models, IC13 for reappraisal and IC7 for suppression usage
| Target | Features | Model | R2 | RMSE | MSE | MAE | MLS | NoL | LR |
|---|---|---|---|---|---|---|---|---|---|
| RS (ERQ) | SS,STA1S, IC13, Tei SS | BT | 0.32 | 0.8489 | 0.7206 | 0.6786 | 8 | 30 | 0.1 |
| SS (ERQ) | RS,IC7, STA1S, Tei SS | BT | 0.16 | 1.0315 | 0.9522 | 0.764 | 8 | 30 | 0.1 |
| SS (ERQ) | RS, IC8, STA1 S, Tei SS | BT | 0.12 | 1.0068 | 1.0136 | 0.7931 | 8 | 30 | 0.1 |
We additionally report the IC8 as the second winning model for suppression as confirmed by stepwise regression.
RS = reappraisal score; SS = suppression score; R2 = coefficient of determination; RMSE = standard deviation of the residuals; MSE = mean squared error; MAE = mean absolute error; MLS = minimum leaf size; NoL = number of learners; LR = learning rate; BT = boosted tree; Tei SS = Tei subscales; STAI S = STAI score.