Table 2 Absolute (and relative) frequencies of all NE classes in each part of the JNLPBA dataset
From: Various criteria in the evaluation of biomedical named entity recognition
| Â | Protein | DNA | RNA | Cell Type | Cell Line | All |
|---|---|---|---|---|---|---|
Training Set | 30,269 (59.0) | 9,533 (18.6) | 951 (1.9) | 6,718 (13.1) | 3,830 (7.5) | 51,301 (100) |
Test Set | 5,067 (58.5) | 1,056 (12.2) | 118 (1.4) | 1,921 (22.2) | 500 (5.8) | 8,662 (100) |