Table 5.
Experimental results achieved on baseline corpora.a
| GSC+ |
EHR |
|||||
|---|---|---|---|---|---|---|
| Method | P | R | F1 | P | R | F1 |
| Exact matching | 94.2 | 42.6 | 58.6 | 99.8 | 58.0 | 73.4 |
| PhenoTagger | 67.3 | 66.3 | 66.8 | 68.3 | 67.9 | 68.1 |
| PhenoTagger (full text) | 77.0 | 67.9 | 72.2 | 59.9 | 70.0 | 64.6 |
| ClinPheno | 63.7 | 65.3 | 64.5 | 60.3 | 62.7 | 61.4 |
| ClinPheno (full text) | 73.2 | 36.0 | 48.3 | 45.8 | 52.5 | 48.9 |
| Doc2HPO | 80.5 | 49.8 | 61.5 | |||
| Monarch Annotator | 82.3 | 50.2 | 62.3 | |||
| NCBO Annotator | 66.0 | 49.1 | 56.3 | |||
a
Bolded values represent the best F1 in class, while italics denote the best recall.