Table 2 Coefficients of selected variables

From: Predicting the naturalistic course of depression from a wide range of clinical, psychological, and biological data: a machine learning approach

A:

Presence of a unipolar depression diagnosis at follow-up

   

Rank a

 

β b

r pb c

    
 

(Intercept)

−0.03

    

1

IDS scored

0.39

0.25

    

2

Conscientiousness

−0.33

−0.19

    

3

Extraversion

−0.04

−0.16

    

4

Neuroticism

−0.06

0.16

    

5

MDD criteriae

0.1

0.14

    

6

Dysthymia lifetime

−0.13

0.15

    

7

Dysthymia 1mf

0.19

0.16

    

8

Dysthymia

0.2

0.15

    

9

Mild recurrent MDD

−0.11

−0.13

    

B:

Course trajectories

Remitted

 

Improved

 

Chronic

 

Rank a

 

β b

r pb c

β b

r pb c

β b

r pb c

 

(Intercept)

0.31

0.09

-0.4

1

IDS scored

−0.31

−0.29

0.12

0.16

0.19

0.16

2

Conscientiousness

0.13

0.16

−0.08

−0.11

−0.04

−0.07

3

Extraversion

0.09

0.2

−0.05

−0.12

−0.04

−0.11

4

Suicidality

−0.1

−0.15

0.1

0.11

0

0.05

5

Dysthymia lifetimef

0.14

−0.16

−0.04

0.02

−0.1

0.16

6

Dysthymia 12mf

−0.04

−0.18

−0.04

0.04

0.09

0.17

7

Dysthymia 6mf

0.24

−0.18

−0.04

0.04

−0.2

0.17

8

Dysthymia 1mf

−0.41

−0.2

0.15

0.06

0.26

0.18

9

Dysthymia

−0.16

−0.16

−0.05

0.02

0.22

0.16

  1. aFeatures are ranked based on order of selection by the stability selection approach
  2. bCoefficients of the logistic regression models. In the case of a multi-class problem (table B), coefficients of each of the binary regressions are shown. However, the direction and a magnitude of coefficients are hard to interpret due to a collinearity problem
  3. cUnivariate (point biserial) correlation coefficients showing the relationship of individual variable with different course groups
  4. dIDS, inventory of depressive symptomatology
  5. eNumber of DSM-IV diagnostic criteria met for a diagnosis of major depressive disorder (MDD)
  6. fRecency of dysthymia in months