Table 1 Cross validated (32 runs of 3:2 train-test split) linear SVM and random forest performance on original datasets, RST transformed data sets, LDA transformed data sets and PCA transformed data sets

Otto group

 RandomForest

93 (original)

1.60 ± 0.03

80

2.12 ± 0.04

1.69 ± 0.03

68

3.44 ± 0.08

1.71 ± 0.03

58

3.39 ± 0.08

1.69 ± 0.02

49

3.36 ± 0.08

1.74 ± 0.02

41

3.36 ± 0.15

1.72 ± 0.03

34

3.25 ± 0.18

1.74 ± 0.03

8

2.12 ± 0.06

 LinearSVM

93 (original)

0.96 ± 0.04

80

0.87 ± 0.04

0.78 ± 0.01

68

0.77 ± 0.05

0.79 ± 0.03

58

0.70 ± 0.01

0.80 ± 0.03

49

0.70 ± 0.01

0.81 ± 0.02

41

0.70 ± 0.01

0.84 ± 0.03

34

0.70 ± 0.01

0.88 ± 0.03

8

0.94 ± 0.02

mnist

 RandomForest

784 (original)

0.46 ± 0.00

684

1.24 ± 0.04

1.55 ± 0.04

597

1.86 ± 0.04

1.50 ± 0.04

454

1.85 ± 0.05

1.53 ± 0.07

396

1.87 ± 0.03

1.46 ± 0.03

345

1.98 ± 0.14

1.33 ± 0.02

288

1.26 ± 0.20

1.29 ± 0.03

9

1.04 ± 0.02

 LinearSVM

784 (original)

1.85 ± 0.10

684

1.17 ± 0.06

1.33 ± 0.05

597

0.60 ± 0.09

1.33 ± 0.04

521

0.68 ± 0.13

1.37 ± 0.11

396

0.71 ± 0.07

1.35 ± 0.07

345

0.55 ± 0.09

1.30 ± 0.06

288

0.41 ± 0.11

1.37 ± 0.07

9

0.52 ± 0.01

Olivetti faces

 RandomForest

4096 (original)

4.29 ± 1.09

720

6.45 ± 0.58

9.58 ± 1.43

509

3.67 ± 0.40

8.24 ± 0.38

430

3.32 ± 0.53

8.04 ± 0.75

375

3.62 ± 0.39

8.13 ± 1.05

327

4.42 ± 0.34

7.29 ± 0.52

288

3.96 ± 0.83

7.05 ± 0.83

39

4.15 ± 0.26

 LinearSVM

4096 (original)

1.98 ± 0.04

720

1.21 ± 0.05

1.51 ± 0.09

509

1.07 ± 0.04

1.50 ± 0.04

430

1.07 ± 0.04

1.53 ± 0.08

375

1.07 ± 0.04

1.51 ± 0.05

327

1.07 ± 0.04

1.52 ± 0.09

285

1.07 ± 0.04

1.52 ± 0.06

39

1.53 ± 0.07