Adapting transformer-based language models for heart disease detection and risk factors extraction

Journal of Big Data

Table 6 RoBERTa Evaluation Metrics for each risk factor indicator

Risk factor	Indicator	Precision	Recall	F1-score	Support
CAD	Mention	0.82	0.79	0.81	228
	Event	0.72	0.70	0.71	278
	Test	0.00	0.00	0.00	70
	Symptom	0.90	0.39	0.54	95
Diabetes	Mention	0.92	0.99	0.95	549
	A1c	0.89	0.79	0.84	71
	Glucose	0.00	0.00	0.00	40
Hypertension	Mention	0.99	0.97	0.98	525
	High bp	0.93	0.99	0.96	210
OBESE	Mention	0.95	1.00	0.97	115
	Obese_BMI	0.00	0.00	0.00	6
Hyperlipidemia	Mention	0.87	0.99	0.93	232
	High chol.	0.00	0.00	0.00	7
	High LDL	0.50	0.10	0.17	30
Smoker	Smoker_never	0.80	0.72	0.76	114
	Smoker_ever	0.00	0.00	0.00	3
	Smoker_current	0.00	0.00	0.00	37
	Smoker_unknown	0.93	0.90	0.91	202
	Smoker_past	0.85	0.59	0.69	121
Medication		0.87	0.74	0.80	7329
Family history		0.2000	0.0769	0.1111	13
Weighted average		0.9390	0.9427	0.9394	42946
Macro average		0.4195	0.4037	0.3915	42946
Micro average				0.9427	42946
Accuracy				0.9427	42946