Classifying the lifestyle status for Alzheimer’s disease from clinical notes using deep learning with weak supervision

Table 2 Comparison of results for models for the physical activity case study

Model	Weighted avg			Physical active			Physical inactivity
Model	Precision	Recall	F1	Precision	Recall	F1	Precision	Recall	F1
Rule-based	0.88	0.53	0.62	0.87	0.32	0.47	0.89	0.85	0.87
Logistic regression	0.91	0.89	0.89	1.00	0.81	0.90	0.77	1.00	0.87
Random forest	0.89	0.88	0.88	0.95	0.85	0.89	0.80	0.93	0.86
SVM	0.91	0.89	0.89	1.00	0.82	0.90	0.79	1.00	0.88
BERT base	0.90	0.90	0.89	0.89	0.94	0.92	0.91	0.82	0.86
Bio BERT	0.89	0.89	0.89	0.91	0.92	0.91	0.88	0.85	0.60
PubMed BERT (Abs)	0.90	0.89	0.89	0.89	0.95	0.92	0.92	0.80	0.85
PubMed BERT (Abs + Ft)	0.88	0.88	0.87	0.86	0.95	0.90	0.91	0.76	0.82
Bio-clinical BERT	0.91	0.91	0.91	0.91	0.95	0.93	0.91	0.86	0.89
UMLS BERT	0.93	0.93	0.92	0.91	0.98	0.94	0.96	0.84	0.89

ISSN: 1472-6947