Skip to main content

Table 1 Performances of traditional machine learning models in terms of class-specific recall, precision and F1-scores, and overall accuracy

From: Text classification models for the automatic detection of nonmedical prescription medication use from social media

Classification algorithm

Precision

Recall

F1-score

Accuracy (%)

A

C

M

U

A

C

M

U

A

C

M

U

Gaussian NB

0.23

0.53

0.84

0.20

0.75

0.43

0.23

0.71

0.36

0.48

0.36

0.32

38.6

SVM (Setting 1)

0.60

0.70

0.74

0.80

0.31

0.62

0.89

0.74

0.41

0.66

0.81

0.77

72.3

SVM (Setting 2)

0.49

0.65

0.78

0.84

0.36

0.70

0.82

0.71

0.41

0.68

0.80

0.77

71.0

RF (n = 100)

0.60

0.65

0.75

0.84

0.20

0.67

0.89

0.72

0.30

0.66

0.81

0.78

71.4

Shallow NN [[32, 16, 8]]

0.43

0.67

0.77

0.72

0.46

0.62

0.78

0.72

0.44

0.64

0.77

0.72

68.4

KNN (3)

0.31

0.57

0.66

0.76

0.34

0.35

0.79

0.57

0.32

0.43

0.72

0.65

58.8

  1. Best F1-score on the A class is shown in bold