Table 9 NER performance before and after typo correction in the SPR dataset
From: MLM-based typographical error correction of unstructured medical texts for named entity recognition
NER | Precision | Recall | f1-score | Support | |||
|---|---|---|---|---|---|---|---|
Typos | Typo correction | Typos | Typo correction | Typos | Typo correction | ||
B-ORGAN | 0.97 | 1.00 | 0.31 | 0.82 | 0.47 | 0.90 | 1022 |
I-ORGAN | 1.00 | 1.00 | 0.37 | 0.73 | 0.54 | 0.84 | 196 |
B-LOCATION | 1.00 | 1.00 | 0.25 | 0.79 | 0.41 | 0.88 | 625 |
I-LOCATION | 1.00 | 1.00 | 0.26 | 0.81 | 0.41 | 0.89 | 1338 |
B-OPNAME | 0.90 | 0.98 | 0.54 | 0.97 | 0.68 | 0.97 | 822 |
I-OPNAME | 1.00 | 1.00 | 0.46 | 0.96 | 0.63 | 0.98 | 803 |
B-HISTOLOGIC DIAGNOSIS | 1.00 | 1.00 | 0.83 | 0.99 | 0.91 | 0.99 | 70 |
I-HISTOLOGIC DIAGNOSIS | 1.00 | 1.00 | 0.90 | 1.00 | 0.95 | 1.00 | 114 |
B-TUMOR_SIZE | 1.00 | 1.00 | 0.99 | 0.99 | 1.00 | 1.00 | 222 |
| Â | Typos | Typo correction | Â | ||||
Accuracy | 0.40 | 0.87 | 2047 | ||||
f1-score | 0.60 | 0.85 | |||||