Skip to main content

Table 3 Comparison of results for models for the excessive diet case study

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

Model

Weighted avg

Normal diet

High calorie diet

Precision

Recall

F1

Precision

Recall

F1

Precision

Recall

F1

Rule-based

0.91

0.86

0.87

0.52

0.92

0.67

0.94

0.94

0.94

Logistic regression

0.88

0.85

0.85

0.50

0.83

0.63

1.00

0.94

0.97

Random forest

0.89

0.86

0.86

0.52

0.89

0.66

1.00

0.94

0.97

SVM

0.88

0.85

0.85

0.50

0.83

0.63

1.00

0.94

0.97

BERT base

0.91

0.91

0.91

0.66

0.75

0.70

1.00

0.98

0.99

Bio BERT

0.92

0.92

0.92

0.71

0.78

0.74

1.00

0.98

0.99

PubMed BERT (Abs)

0.91

0.90

0.90

0.59

0.81

0.68

1.00

1.00

1.00

PubMed BERT (Abs + Ft)

0.90

0.90

0.90

0.63

0.64

0.62

1.00

1.00

1.00

Bio-clinical BERT

0.93

0.93

0.93

0.73

0.75

0.73

1.00

1.00

1.00

UMLS BERT

0.92

0.92

0.92

0.72

0.69

0.70

1.00

1.00

1.00

 

High fat diet

High salt diet

Nonspecific abnormal

Precision

Recall

F1

Precision

Recall

F1

Precision

Recall

F1

Rule-based

1.00

0.90

0.95

1.00

0.95

0.97

0.92

0.61

0.73

Logistic regression

0.95

0.95

0.95

1.00

1.00

1.00

0.82

0.50

0.62

Random forest

0.95

0.95

0.95

0.98

1.00

0.99

0.87

0.50

0.64

SVM

0.95

0.95

0.95

1.00

1.00

1.00

0.81

0.50

0.62

BERT base

0.98

1.00

0.99

1.00

1.00

1.00

0.82

0.74

0.78

Bio BERT

0.98

0.98

0.99

1.00

1.00

1.00

0.85

0.80

0.82

PubMed BERT (Abs)

1.00

0.98

0.99

1.00

1.00

1.00

0.84

0.65

0.73

PubMed BERT (Abs + Ft)

1.00

0.98

0.99

1.00

1.00

1.00

0.77

0.76

0.76

Bio-clinical BERT

1.00

0.98

0.99

1.00

1.00

1.00

0.85

0.83

0.84

UMLS BERT

1.00

1.00

1.00

1.00

1.00

1.00

0.81

0.81

0.81

  1. *Bold numbers indicate best performance in each column