Research Article
Phenonizer: A Fine-Grained Phenotypic Named Entity Recognizer for Chinese Clinical Texts
Table 5
Comparison of degraded models for COVID-19 EMRs.
| Degraded models | Precision | Recall | F1-score |
| BiLSTM-CRF | 0.8660 | 0.8515 | 0.8587 | GloVeWiki-BiLSTM-CRF | 0.8620 | 0.8468 | 0.8544 | GloVeMedical-BiLSTM-CRF | 0.8691 | 0.8481 | 0.8585 | W2VWiki-BiLSTM-CRF | 0.8605 | 0.8629 | 0.8617 | W2VMedical-BiLSTM-CRF | 0.8624 | 0.8701 | 0.8662 | BERT-CRF | 0.8664 | 0.8916 | 0.8788 | Phenonizer | 0.8767 | 0.8852 | 0.8809 |
|
|