Deep learning with language models improves named entity recognition for PharmaCoNER

BMC Bioinformatics

Table 4 Performance comparison on the PharmaCoNER dataset

Method	P (%)	R (%)	F1 (%)
Baseline-Glove [32]	83.26	81.00	82.11
Baseline-Med [32]	87.02	83.71	85.34
Sun et al. [21]	90.46	88.06	89.24
Stoeckel et al. [19]	90.79	90.30	90.52
Xiong et al. [18]	91.23	90.88	91.05
Our method (BioBERTv1.1(+P,Cased))	92.44	91.59	92.01

ISSN: 1471-2105