Table 1.
Statistics of n2c2 dataset in intra- and inter-sentence relations for the training and development sets
| Training | Development | |||
|---|---|---|---|---|
| Total sentences | 44 475 | 11 520 | ||
| Sentences with >1 entity | 7125 | 1907 | ||
| Sentences with 1 entity | 1835 | 401 | ||
| Sentences with no entities | 35 515 | 9212 | ||
| Sentences with 1 pair | 1672 | 409 | ||
| Inter | Intra | Inter | Intra | |
| Number of positive relations | 1994 | 26591 | 570 | 7119 |
| Strength-Drug | 36 | 5276 | 13 | 1373 |
| Dosage-Drug | 107 | 3192 | 33 | 888 |
| Duration-Drug | 29 | 489 | 4 | 120 |
| Frequency-Drug | 158 | 4828 | 53 | 1259 |
| Form-Drug | 123 | 5060 | 74 | 1358 |
| Route-Drug | 107 | 4220 | 35 | 1173 |
| Reason-Drug | 1239 | 2830 | 307 | 783 |
| ADE-Drug | 195 | 696 | 51 | 165 |
| Negative relations | 78 471 | 39 850 | 20 050 | 9183 |
| Duplicate relations | 19 | 9 | ||
ADE: adverse drug event.