Skip to main content

Table 5 The effect of variables on prediction of deaths from cardiovascular diseases based on variable importance measure

From: Optimizing cardiovascular disease mortality prediction: a super learner approach in the tehran lipid and glucose study

Variable

Cox-PH (%)

GBM (%)

SVM (%)

SL (%)

Age

34.80

38.98

49.06

31.59

Waist-to-hip ratio

21.95

2.09

24.79

8.83

Smoking status

15.78

7.35

15.44

8.91

Systolic blood pressure

5.59

11.74

1.59

10.07

Diastolic blood pressure

5.01

3.97

1.26

8.29

Fasting blood sugar

4.13

17.52

1.17

7.31

Body mass index

3.09

1.52

0.71

5.77

Type 2 diabetes mellitus

1.99

3.54

0.85

3.20

Aspirin

1.06

5.75

1.08

3.64

Physical Activity

0.98

2.17

0.54

1.97

Cholesterol

0.97

1.32

0.69

1.41

Triglyceride

0.91

0.96

0.61

1.01

High-density lipoprotein

0.87

1.15

0.66

1.07

Sex

0.83

0.84

0.33

0.94

Lipid Drug

0.75

0.46

< 0.001

0.19

Family History of Stroke

0.54

< 0.001

0.74

2.93

Marital Status

0.38

< 0.001

0.23

1.29

Anti-Hypertension Drug

0.22

0.63

0.01

0.15

Education

0.13

< 0.001

0.21

0.86

Beta-Blockers

0.09

< 0.001

0.01

0.34

Corticosteroid

0.08

< 0.001

0.01

0.23