Skip to main content

Table 7 Average MSEs and their standard deviation obtained by DFRT and RF on the training and test sets for the real world and synthetic datasets

From: Building efficient fuzzy regression trees for large scale and high dimensional problems

Prob.

DFRT

RF

Training ± s.d.

Test ± s.d.

Training ± s.d.

Test ± s.d.

Window2

24.3263 ± 0.18

24.3978 ± 0.90

22.9981 ± 3.02

29.9958 ± 5.72

Window3

24.2976 ± 0.18

24.3853 ± 0.91

23.1057 ± 3.43

29.0574 ± 5.35

Window4

24.2841 ± 0.18

24.3730 ± 0.91

21.3935 ± 1.31

28.9095 ± 5.13

Window5

24.2833 ± 0.18

24.3732 ± 0.91

22.1959 ± 2.85

28.8784 ± 4.95

Window6

24.2834 ± 0.18

24.3733 ± 0.91

22.5899 ± 3.53

29.0761 ± 5.52

Window7

24.2836 ± 0.18

24.3738 ± 0.91

21.9378 ± 2.56

29.7301 ± 4.78

Window8

24.2841 ± 0.18

24.3742 ± 0.91

21.0925 ± 2.64

31.1765 ± 5.26

Window9

24.2807 ± 0.18

24.3689 ± 0.91

21.8789 ± 2.82

28.9022 ± 4.94

\(f_1\)

15.0433 ± 0.03

16.6537 ± 0.18

17.6090 ± 0.48

15.0783 ± 0.84

\(f_{19}\)

6.5671 ± 0.02

7.3154 ± 0.11

6.6059 ± 0.08

6.8696 ± 0.09

\(f_2\)

31.3784 ± 0.02

32.0971 ± 0.14

32.2631 ± 0.28

29.1347 ± 0.75

\(f_4\)

428.8752 ± 0.47

429.8814 ± 3.39

426.1867 ± 0.74

427.2893 ± 3.40

  1. The best results for each dataset have been emphasized in italic font
  2. The results have to be multiplied by \(10^{21}\), \(10^{22}\), \(10^4\) and \(10^{31}\) for \(f_1\), \(f_{19}\), \(f_2\) and \(f_4\) respectively