Using natural language processing methods to classify use status of dietary supplements in clinical notes

Table 7 Comparison between the rule-based and the machine learning-based classifiers regarding four use statuses

Status	Number	TP^a for RB^a	TP for ML^a	RB (+)^a ML (+)	RB (+) ML (−)^a	RB (−) ML (+)	RB (−) ML (−)
C	233	212	222	209	3	13	8
D	166	132	148	128	4	20	14
S	178	151	163	147	4	16	12
U	173	166	144	142	24	2	4

^aTP: true positive; RB: rule-based classifier; ML: machine learning-based classifier; (+): correctly classified; (−): falsely classified

Status	Number	TP^a for RB^a	TP for ML^a	RB (+)^a ML (+)	RB (+) ML (−)^a	RB (−) ML (+)	RB (−) ML (−)
C	233	212	222	209	3	13	8
D	166	132	148	128	4	20	14
S	178	151	163	147	4	16	12
U	173	166	144	142	24	2	4

Status	Number	TP^a for RB^a	TP for ML^a	RB (+)^a ML (+)	RB (+) ML (−)^a	RB (−) ML (+)	RB (−) ML (−)
C	233	212	222	209	3	13	8
D	166	132	148	128	4	20	14
S	178	151	163	147	4	16	12
U	173	166	144	142	24	2	4

Status	Number	TP^a for RB^a	TP for ML^a	RB (+)^a ML (+)	RB (+) ML (−)^a	RB (−) ML (+)	RB (−) ML (−)
C	233	212	222	209	3	13	8
D	166	132	148	128	4	20	14
S	178	151	163	147	4	16	12
U	173	166	144	142	24	2	4