Multi-task learning for Chinese clinical named entity recognition with external knowledge

Table 9 Performances of the networks with and without multi-task learning on the two datasets

Dataset	Method	Precision	Recall	F-measure
CCKS2017	Single-task for NES	92.06	91.55	91.80
	Single-task for NER	91.72	91.59	91.65
	Multi-task for NER	91.91	91.78	91.84
FCCd	Single-task for NES	87.24	86.81	87.02
	Single-task for NER	87.03	86.60	86.81
	Multi-task for NER	87.32	86.79	87.05

ISSN: 1472-6947