Research Article
Research on Named Entity Recognition of Electronic Medical Records Based on RoBERTa and Radical-Level Feature
Table 3
Comparison of three pretraining models on CCKS2019.
| Model | Precision | Recall | F1 score |
| BERT+CRF | 81.82 | 79.33 | 80.56 | ALBERT+CRF | 81.91 | 77.68 | 79.74 | RoBERTa-wwm-ext+CRF | 81.83 | 79.33 | 80.55 | RoBERTa-wwm-ext-large+CRF | 82.29 | 79.69 | 80.97 |
|
|