Table 4 NER performance comparison results of BERT and bi-LSTM-CRF on the Exobrain dataset
From: Korean clinical entity recognition from diagnosis text using BERT
| Â | BERT (IOB) | bi-LSTM-CRF (IOB) | ||
|---|---|---|---|---|
| Â | macro-avg | micro-avg | macro-avg | micro-avg |
Precision | 0.93 | 0.93 | 0.90 | 0.89 |
Recall | 0.94 | 0.94 | 0.87 | 0.86 |
F1 | 0.94 | 0.93 | 0.89 | 0.88 |