Skip to main content

Table 1 Summary variables used to train machine learning algorithms

From: Using alcohol consumption diary data from an internet intervention for outcome and predictive modeling: a validation and machine learning study

Variable name

Description

Mean

Median

Max

Min

SD

Abs.diff

Absolute difference first-last reported drink

−0.39

0

12

−17

3.14

Avg.drinks

Average reported drinks

3.26

3

15

0

2.42

Entries

Total number of entries

10.97

8

48

1

9.73

intercept

Intercept of trajectory

0.05

0.22

4.79

−2.73

1.13

IQR.drinks

Inter-quartile range of drinks

1.69

1

13.5

0

1.89

Max.drinks

Maximum reported drinks

6.05

6

22

0

3.98

Median.drinks

Median reported drinks

2.92

3

15

0

2.62

Min.drinks

Minimum reported drinks

1.67

1

15

0

2.24

n.binge

Number of binge drinking entries

0.72

0

20

0

1.71

n.heavy

Number of heavy drinking entries

2.35

1

35

0

3.32

n.light

Number of light drinking entries

7.9

4

48

0

8.79

Perc.binge

Percentage binge drinking entries

0.09

0

1

0

0.2

Perc.heavy

Percentage of heavy drinking entries

0.26

0.17

1

0

0.3

Perc.light

Percentage of light drinking entries

0.65

0.75

1

0

0.35

Range.drinks

Range of reported drinks

4.39

4

22

0

4.15

Rel.diff

Relative difference first-last reported drinks

−0.06

0

2.5

−4.5

0.61

slope

Slope of trajectory

−0.01

−0.01

9.21

−14

2.34

Sum.drinks

Total sum of reported drinks

28.29

16

208

0

32.23

  1. 1Bing-drinking defined as > 6 for women, > 8 for men
  2. 2Heavy drinking defined as > 3 and < 7 for women, > 4 and < 9 for men
  3. 3Light drinking defined as < 4 drinks for women, < 5 for men