Table 5 Classification accuracy of the proposed ELM-based ensemble methods under a bagging framework.

From: M-estimation activation functions for high-performance extreme learning machine ensemble classification

Activation functions

Sat image ±

Emails Spamdexing ±

Breast Cancer ±

Musk ±

Iris ±

Sigmoid

88.43

0.64621

89.33333

1.527525

73.8

19.21458

83.0714

3.791981

96.98

2.706681

Tan-Sig

88.929

0.47463

93

2

74.4

11.14899

83.0714

3.791981

97.4

2.364706

Sine

89

0.67936

94.33333

1.527525

83.8

12.65701

88.1428

4.203609

96.86

2.703059

Cosine

89.142

0.53452

77.33333

1.527525

73.8

13.04607

89.6428

2.468483

95.71429

2.757607

Bentidle

87.357

0.84189

93.66667

0.57735

79.8

7.155418

80.4285

3.030976

96.71429

2.840059

RAF

89.571

0.6462

77

1

86.2

3.03315

89.2142

3.490175

97

2.287087

Proposed 1

89

0.5547

95.33333

2.516611

79.6

11.99166

86

5.43493

96.78571

2.913591

Proposed 2

88.714

0.61124

93.66667

1.154701

86.2

5.932959

87.5714

2.765565

96.42857

1.741542

Proposed 3

75

0.67936

92.33333

0.57735

95.4

2.073644

74.642

2.405351

85

4.574175

Proposed 4

90.71

0.4688

90.66667

1.527525

76

6.892024

86.857

2.24832

96.14286

2.741621

Proposed 5

88.571

0.51355

95 0.0000

1

77.6

18.18791

86.3571

3.103525

96.57143

2.440501

Proposed 6

87.643

0.74494

95 0.0000

1

79.6

13.57571

77.3571

3.607836

97.571

1.603567

Proposed 7

72.214

3.16661

72.33333

2.081666

72.2

5.263079

80.0714

3.338915

95

2.112235

Proposed 8

84.428

0.6462

91.66667

1.154701

87.8

4.266146

81.2142

3.26

95.357

3.028255

Proposed 9

88.42857

0.75592

95.33333

1.564234

77.2

11.2783

89.3371

1.87934

97.5

2.441626

Proposed 10

81.5

0.94053

71.66667

0.57735

53.8

13.40522

86.6592

2.88736

95

4.420233

Proposed 11

89.142

0.66299

95.33333

1.154701

91

4.358899

87.0673

5.76564

97.714

1.540658

Back to article page