Table 4.
Number of entity (mention and identifier) and relation annotations in the BioRED corpus, the IAA and the distribution between the training, development and test sets. The parenthesized numbers are the unique entities linked with concept identifiers.
| Annotation s | Training | Dev | Tests | Total | IAA | |
|---|---|---|---|---|---|---|
| Document | 400 | 100 | 100 | 600 | - | |
| Entity (ID) | All | 13 351 (2708) | 3533 (956) | 3535 (982) | 20 419 (3869) | 97.01% |
| Gene | 4430 (1141) | 1087 (368) | 1180 (399) | 6697 (1643) | 97.35% | |
| Disease | 3646 (576) | 982 (244) | 917 (244) | 5545 (778) | 96.06% | |
| Chemical | 2853 (486) | 822 (184) | 754 (170) | 4429 (651) | 96.12% | |
| Variant | 890 (420) | 250 (135) | 241 (137) | 1381 (678) | 97.79% | |
| Species | 1429 (37) | 370 (13) | 393 (11) | 2192 (47) | 99.43% | |
| Cell Line | 103 (48) | 22 (12) | 50 (21) | 175 (72) | 99.68% | |
| Relation | 4178 | 1162 | 1163 | 6503 | 77.91% | |
| Relation pair with novelty findings | 2838 | 835 | 859 | 4532 | 85.01% | |