Table 5 Statistical metrics of the B-WKELM-R-GBO model and four other standalone ML models.

From: Development of a novel modeling framework based on weighted kernel extreme learning machine and ridge regression for streamflow forecasting

 

Standalone model

 

R

RMSE

MAPE

NSE

IA

U95%

Q(t + 3)

K = 0

B-WKELM-R-GBO

Train

0.817

1.979

45.309

0.657

0.872

5.485

Test

0.897

1.504

69.562

0.788

0.929

4.170

B-Ridge-GBO

Train

0.729

2.310

77.638

0.532

0.819

6.404

Test

0.873

1.674

110.511

0.738

0.906

4.640

B-GRNN-GBO

Train

0.724

2.353

94.431

0.515

0.795

6.515

Test

0.818

1.916

169.945

0.657

0.873

5.308

B-ENET-GBO

Train

0.730

2.309

77.123

0.533

0.820

6.401

Test

0.874

1.667

109.984

0.740

0.907

4.621

B-LGBM-GBO

Train

0.797

2.043

66.386

0.634

0.869

5.664

Test

0.872

1.608

95.906

0.758

0.923

4.457

Q(t + 7)

K = 0

B-WKELM-R-GBO

Train

0.598

2.801

89.961

0.313

0.589

7.745

Test

0.727

2.397

140.414

0.463

0.726

6.639

B-Ridge-GBO

Train

0.442

3.031

150.187

0.195

0.493

8.403

Test

0.681

2.532

221.344

0.401

0.668

7.016

B-GRNN-GBO

Train

0.558

2.840

159.996

0.293

0.581

7.872

Test

0.613

2.657

282.758

0.340

0.630

7.355

B-ENET-GBO

Train

0.448

3.021

145.311

0.200

0.511

8.375

Test

0.684

2.504

213.889

0.414

0.684

6.939

B-LGBM-GBO

Train

0.653

2.922

181.653

0.252

0.456

8.099

Test

0.525

2.873

282.853

0.229

0.485

7.950

Q(t + 14)

K = 0

B-WKELM-R-GBO

Train

0.359

3.192

115.838

0.109

0.384

8.807

Test

0.436

2.963

190.335

0.181

0.462

8.206

B-Ridge-GBO

Train

0.300

3.226

191.862

0.090

0.317

8.941

Test

0.413

2.999

274.214

0.161

0.402

8.312

B-GRNN-GBO

Train

0.333

3.195

193.917

0.107

0.332

8.855

Test

0.348

3.077

289.754

0.117

0.330

8.529

B-ENET-GBO

Train

0.300

3.226

191.906

0.090

0.317

8.941

Test

0.413

3.000

274.280

0.161

0.402

8.312

B-LGBM-GBO

Train

0.370

3.146

137.049

0.134

0.382

8.720

Test

0.399

3.007

189.485

0.156

0.446

8.338

  1. Significant values are given in bold.
Back to article page