Named Entity Recognition in Chinese Medical Literature Using Pretraining Models

Table 5

Recognition results of pretraining models and deep learning models.

	P (%)	R (%)	F (%)
Deep learning models
BiGRU	89.431	81.842	85.443
BiGRU-CRF	88.463	84.332	86.341
BiLSTM	89.511	82.251	85.700
BiLSTM-CRF	89.113	84.983	86.992
Pretraining models
BERT	90.968	91.048	91.008
BERT-WWM	91.023	91.108	91.065
BERT-WWM-EXT	91.059	90.996	91.027
ERNIE	90.659	90.527	90.593
ERNIE-tiny	89.466	89.348	89.407
RoBERTa	91.164	91.254	91.209