Table 3.
Results from simulations on the corpus of video vignettes used by Cartmill et al. (2013).
| Precision | Recall | F1 | |
|---|---|---|---|
| Propose/Verify | 0.03 | 0.42 | 0.06 (95% CI = 0.057–0.061) |
| Fazly et al. 2010 | 0.47 | 0.37 | 0.41 |
| Modified Fazly | 0.17 | 0.71 | 0.27 |
| Pursuit | 0.59 | 0.29 | 0.39 (95% CI = 0.378–0.394) |