A semi-supervised approach for extracting TCM clinical terms based on feature words

Table 2 Extraction result for TCM clinical terms in the test dataset

Type	Character Vector	Feature Word Categories	Corpus	Precision	Recall	F1
Supervised	WIKI	0	Train_2w +Dev_3.5 k +Test_5.1 k	0.6831	0.7053	0.6941
Semi-Supervised	WIKI	0	Train_2w_10w +Dev_3.5 k +Test_5.1 k	0.7782	0.7294	0.7530
	TCM	0		0.7474	0.7262	0.7339
	0.3TCM + 0.7WIKI	0		0.7893	0.7294	0.7561
	0.7TCM + 0.3WIKI	0		0.7503	0.7335	0.7298
	0.5TCM + 0.5WIKI	0		0.7935	0.7228	0.7585
	0.5TCM + 0.5WIKI	4		0.7889	0.7525	0.7703
	0.5TCM + 0.5WIKI	39		0.7932	0.7723	07826
	0.5TCM + 0.5WIKI	8		0.7756	0.7987	0.7870

ISSN: 1472-6947