Table 6 Performance of the ensemble extractor on the CEM task at various confidence thresholds.
From: Chemical entity extraction using CRF and an ensemble of extractors
Dataset | Threshold | Precision | Recall | F-Measure |
|---|---|---|---|---|
Dev | 0.01 | 0.31543 | 0.8924 | 0.46611 |
Dev | 0.24 | 0.67406 | 0.73650 | 0.70390 |
Dev | 0.25 | 0.70871 | 0.71598 | 0.71232 |
Dev | 0.5 | 0.79486 | 0.67544 | 0.7303 |
Dev | 0.7 | 0.87369 | 0.55663 | 0.6800 |
Dev | 0.8 | 0.88315 | 0.52835 | 0.66116 |
Dev | 0.9 | 0.93316 | 0.30973 | 0.46509 |
Train | 0.01 | 0.30711 | 0.88147 | 0.45552 |
Train | 0.25 | 0.66208 | 0.73126 | 0.69495 |
Train | 0.26 | 0.78473 | 0.66680 | 0.72098 |
Train | 0.5 | 0.78473 | 0.6668 | 0.72098 |
Train | 0.6 | 0.86928 | 0.55312 | 0.67607 |
Train | 0.7 | 0.88266 | 0.52568 | 0.65893 |
Train | 0.9 | 0.93386 | 0.31135 | 0.467 |