Table 4 Performance in the IPS task, using tenfold cross-validation on the training set
From: Automating curation using a natural language processing pipeline
Run | File type | Normalizer | Micro-averaged | Macro-averaged | ||||
|---|---|---|---|---|---|---|---|---|
| Â | Â | Â | Precision | Recall | F1 | Precision | Recall | F1 |
1 | Exact | 0.2680 | 0.1712 | 0.2089 | 0.1945 | 0.2162 | 0.1784 | |
2 | HTML | Exact | 0.2552 | 0.1692 | 0.2035 | 0.1840 | 0.2100 | 0.1708 |
3 | Fuzzy | 0.2336 | 0.1766 | 0.2011 | 0.1901 | 0.2211 | 0.1757 | |