Research Article
Named Entity Recognition in Chinese Medical Literature Using Pretraining Models
Table 6
Recognition results of feature extraction and fine-tuning.
| | Pretraining models | P (%) | R (%) | F (%) |
| | BERT (feature extraction) | 90.881 | 90.983 | 90.932 | | BERT (fine-tuning) | 90.968 | 91.048 | 91.008 | | ERNIE (feature extraction) | 90.519 | 90.639 | 90.579 | | ERNIE (fine-tuning) | 90.659 | 90.527 | 90.593 | | RoBERTa (feature extraction) | 91.109 | 91.275 | 91.192 | | RoBERTa (fine-tuning) | 91.164 | 91.254 | 91.209 |
|
|
Values in bold represent the maximum values.
|
