Table 4.
Evaluation of annotations automatically generated by the text mining-assisted workflow against gold standard data
| Exact matching | Relaxed matching | |||||
|---|---|---|---|---|---|---|
| Precision | Recall | F-score | Precision | Recall | F-score | |
| AnatomicalConcept | 0.1923 | 0.7527 | 0.3063 | 0.2814 | 0.9038 | 0.4292 |
| Drug | 0.5861 | 0.2744 | 0.3738 | 0.7921 | 0.6463 | 0.7118 |
| MedicalCondition | 0.0290 | 0.2842 | 0.2868 | 0.3697 | 0.6313 | 0.4663 |
| TestOrMeasure | 0.1425 | 0.0680 | 0.0920 | 0.1914 | 0.1039 | 0.1347 |
| Treatment | 0.3080 | 0.1494 | 0.2012 | 0.4688 | 0.4015 | 0.4325 |
| Micro-average | 0.2670 | 0.2283 | 0.2462 | 0.4050 | 0.5243 | 0.4570 |
| Macro-average | 0.3037 | 0.3057 | 0.3047 | 0.4207 | 0.5374 | 0.4719 |
Results are reported for only nine full-text papers.