Skip to main content

Table 3 Results of all tests conducted with data and machine learning models, with stratified cross-validation

From: Predicting the outcome for COVID-19 patients by applying time series classification to electronic health records

Dataset

Model

COVID-19 wave

Accuracy (%)

Sensitivity (%)

Precision (%)

f1 score (%)

Specificity (%)

Negative predictive value (%)

Independent days of hospitalization

Univariate

First

81.04

85.16

86.28

85.71

72.75

70.89

Second

83.76

93.15

86.69

89.80

52.73

66.97

Multivariate

First

80.66

87.74

83.98

85.81

66.52

73.06

Second

83.51

90.56

86.99

88.71

65.80

73.64

Complete hospitalization history

Univariate

First

87.55

86.50

94.94

90.52

90.60

76.70

Second

95.98

99.31

95.85

97.55

82.29

96.64

Multivariate

First

94.49

97.75

94.23

95.94

87.96

95.23

Second

95.62

98.65

95.54

97.06

87.43

96.14

  1. The best metrics are bolded and the worst metrics are in italic