Table 2 Comparison of intra-sentence and cross-sentence extraction on the development data
From: Unsupervised inference of implicit biomedical events using context triggers
Gold intra-sentence events only | ||||||
| Â | Gold | Predicted | Correct | Precision | Recall | F1 |
Intra | 165 | 165 | 165 | 100.0 | 100.0 | 100.0 |
Cross | 58 | 0 | 0 | 0.0 | 0.0 | 0.0 |
All | 223 | 165 | 165 | 100.0 | 74.0 | 85.1 |
Gold intra-sentence events + our cross-sentence extraction | ||||||
| Â | Gold | Predicted | Correct | Precision | Recall | F1 |
Intra | 165 | 165 | 165 | 100.0 | 100.0 | 100.0 |
Cross | 58 | 40 | 17 | 42.5 | 29.3 | 34.7 |
All | 223 | 205 | 182 | 88.8 | 81.6 | 85.0 |
Our intra-sentence extraction only | ||||||
| Â | Gold | Predicted | Correct | Precision | Recall | F1 |
Intra | 165 | 325 | 153 | 47.1 | 92.7 | 62.4 |
Cross | 58 | 0 | 0 | 0.0 | 0.0 | 0.0 |
All | 223 | 325 | 153 | 47.1 | 68.6 | 55.8 |
Our intra-sentence & cross-sentence extraction | ||||||
| Â | Gold | Predicted | Correct | Precision | Recall | F1 |
Intra | 165 | 325 | 153 | 47.1 | 92.7 | 62.4 |
Cross | 58 | 47 | 20 | 42.6 | 34.5 | 38.1 |
All | 223 | 372 | 173 | 46.5 | 77.6 | 58.2 |