Table 4 Prediction results for each model.

From: Multi-scale wind speed prediction model based on improved escape algorithm for optimizing time-varying filtering empirical modal decomposition

Pacemaker

Algorithms

R2

MAE

RMSE

One-step

Multi-scale

0.9795

0.5109

0.7541

XLSTM

0.9613

0.7298

1.0359

T-LSTM

0.8785

1.2493

1.8364

Transformer

0.8049

1.6201

2.3277

LSTM

0.9260

0.9592

1.4333

BPNN

0.8931

1.2723

1.7231

ELM

0.8068

1.6547

2.3162

Three-step

Multi-scale

0.9716

0.6295

0.8868

XLSTM

0.9076

1.1061

1.6000

T-LSTM

0.5817

2.4275

3.4055

Transformer

0.5101

2.7783

3.6851

LSTM

0.8139

1.7814

2.2713

BPNN

0.7652

1.9385

2.5513

ELM

0.6608

2.2478

3.0663

Six-step

Multi-scale

0.9504

0.8778

1.1705

XLSTM

0.8199

1.5862

2.2321

T-LSTM

0.7535

1.9369

2.6112

Transformer

0.4943

2.7614

3.7404

LSTM

0.6709

2.2734

3.0175

BPNN

0.6326

2.4751

3.1881

ELM

0.5601

2.5949

3.4887

Nine-step

Multi-scale

0.8592

1.4923

1.9715

XLSTM

0.6451

2.4143

3.1309

T-LSTM

0.6169

2.4977

3.2529

Transformer

0.1656

3.6281

4.8008

LSTM

0.5162

2.8125

3.6555

BPNN

0.4299

2.9634

3.9683

ELM

0.3084

3.2148

4.3707

Twelve-steps

Multi-scale

0.8266

1.6558

2.1899

XLSTM

0.5394

2.7137

3.5688

T-LSTM

0.4656

2.8243

3.8441

Transformer

0.4505

3.0939

3.8980

LSTM

0.4375

2.9399

3.9440

BPNN

0.5188

2.8075

3.6479

ELM

0.3439

3.1376

4.2594

Fifteen-steps

Multi-scale

0.7685

1.9365

2.5327

XLSTM

0.4640

2.9637

3.8545

T-LSTM

0.2175

3.5387

4.6570

Transformer

0.1953

3.6283

4.7226

LSTM

0.3798

3.2676

4.1462

BPNN

0.4945

2.7637

3.7430

ELM

0.2527

3.3090

4.5511

Back to article page