Named Entity Recognition in Chinese Medical Literature Using Pretraining Models

Table 2

Statistics of “A Labelled Chinese Dataset for Diabetes.”

	Training set	Development set	Test set
Disease related
Disease	25197	8399	8399
Reason	2849	950	950
Symptom	3166	1055	1056
Test	28819	9606	9606
Test value	6402	2134	2134
Therapy related
Drug	9946	3315	3315
Frequency	309	103	103
Amount	871	290	290
Method	606	202	203
Treatment	896	298	299
Operation	493	164	164
Side effect	1052	351	350
Common entities
Duration	6543	2181	2180
Anatomy	16866	5622	5622
Level	1333	446	448
Total	105348	35116	35119