Table 4 Recursive Feature Elimination with Cross-Validation using Random Forest Classification results for airport terminal (A) and environmental (B) tests.

From: Machine learning prediction and classification of behavioral selection in a canine olfactory detection program

 

MP

H1

H2

HG

ACT

CR

PP

IP

  

A

3 MO

4.7

–

30.7

86.7

–

–

–

–

  

6 MO

24.7

–

10.7

–

–

–

100.0

47.3

  

9 MO

–

12.7

53.3

–

–

–

64.0

19.3

  

12 MO

48.7

38.7

77.3

28.7

0.0

6.0

51.3

76.0

  
 

Conf

Conc

Respon

Init

Excit

Hear Sens

Body Sens

CR

PP

IP

B

3 MO

100.0

–

–

0.4

–

–

–

–

–

–

6 MO

88.7

–

4.7

13.3

–

–

–

–

–

–

9 MO

20.7

16.0

23.3

30.7

16.7

0.8

12.0

14.0

23.3

93.3

12 MO

63.3

48.9

32.0

32.7

46.7

21.3

22.7

28.7

80.7

47.3

  1. Values indicate the percentage of 250 bootstrap runs the respective traits appeared in, ranging from 0 to 100.
Back to article page