Table 4. Overall performance.

Method	Metric	MACCROBAT	NCBI-Disease	I2b2-2012
BiLSTM-CNN-Char	Precision	84.43	85.24	79.35
	Recall	83.97	83.31	78.11
	F1-score	84.20	84.26	78.73
SciBERT	Precision	78.10	76.88	77.01
	Recall	72.18	74.10	75.18
	F1-score	75.02	75.46	76.08
BlueBERT	Precision	84.04	83.37	81.10
	Recall	81.48	81.39	80.88
	F1-score	82.74	82.37	80.99
ClinicalBERT	Precision	81.01	84.08	80.35
	Recall	79.10	80.11	78.69
	F1-score	80.04	82.05	79.51
BioBERT v1.2	Precision	86.72	85.80	88.00
	Recall	88.31	84.29	86.10
	F1-score	87.51	85.04	87.04
BioEN (our approach)	Precision	92.10	91.68	90.10
	Recall	91.68	88.92	88.98
	F1-score	91.89	90.28	89.54