Skip to main content

Table 4 Importance of each variable in Random Forest analysis

From: What makes a successful species? Traits facilitating survival in altered tropical forests

Variable

Complete

D vs. SD

D vs. NR

D vs. I

I vs. NR

SVL females

6.73

3.41

3.42

2.15

2.08

Adult habitat

4.70

3.23

3.64

0.91

2.37

Sexual dimorphism

9.02

4.00

2.98

2.63

2.59

Clutch size category

6.15

3.49

3.39

4.20

2.63

Egg deposition site

1.13

1.04

0.46

0.50

0.48

Reproductive mode

0.14

0.47

0.07

0.10

0.14

Larval habitat

3.46

2.63

2.10

0.89

1.12

Region

4.54

4.95

2.62

2.11

1.58

Range size

12.52

5.91

7.52

3.27

3.06

Elevational range

8.59

3.37

6.81

2.23

2.87

  1. Importance of each variable per Random Forest analysis (complete data set and different subsets); the four most important variables contributing to the classification are in italics; analysis were performed with ntree = 1000, mtry = 3 and sampsize adjusted based on the smallest sample size for each analysis respectively (R package ‘randomForest’); forest dependency index: D dependent (n = 33), SD slightly dependent (n = 108), NR non-responding (n = 83), I forest independent (n = 19)