De-identification of clinical notes with pseudo-labeling using regular expression rules and pre-trained BERT

Table 7 Result of KoBERT-NER and SciBERT-NER

Training level	Base KoBERT-NER (0)		SciBERT-Base (0)		KoBERT-Large (32,000)		SciBERT-Large (32,000)
Training level	Precision	Recall	Precision	Recall	Precision	Recall	Precision	Recall
DAT	0.49	0.43	0.34	0.05	0.99	0.98	0.98	1.00
PER	0.46	0.40	1.00	0.40	0.89	0.93	1.00	0.79
ORG	0.02	0.47	0.05	0.43	0.69	0.96	0.79	0.87
NUM	0.02	0.77	0.01	0.54	1.00	0.57	1.00	0.57
LOC	0.50	0.11	0.0	0.0	0.81	0.78	1.00	0.67
ETC	0.0	0.0	0.0	0.0	1.00	1.00	1.00	0.54
Total	0.46	0.43	0.35	0.09	0.96	0.92	0.97	0.78

ISSN: 1472-6947