Table 2 Comparison of models on meteorological datasets.

From: Sequence to sequence architecture based on hybrid LSTM global and local encoders approach for meteorological factors forecasting

Baseline

SML2010-Hum

SML2010-outTem

MSE

MAE

MAPE

MSE

MAE

MAPE

NARM

0.427

0.411

2.035

0.382

0.374

1.899

ConvS2S

0.413

0.380

1.913

0.367

0.368

1.862

COPYNET

0.508

0.495

2.438

0.412

0.403

2.024

PGNET

0.472

0.462

2.309

0.413

0.412

2.150

FEDformer

0.419

0.401

1.953

0.401

0.387

1.974

Crossformer

0.398

0.382

1.887

0.326

0.314

1.757

DeepAR

0.433

0.421

1.906

0.390

0.377

1.921

N-BEATS

0.510

0.492

2.317

0.499

0.492

2.178

WRQASA

0.188 a

0.205

1.143

0.301 b

0.308

1.729

STConvS2S

0.309

0.281

1.820

0.342

0.338

1.958

SFA-LSTM

0.271

0.265

1.784

0.330

0.329

1.826

H-LSTM-GLE

0.196

0.192a

1.071a

0.312

0.302b

1.617b

  1. aBold indicates the optimal value of SML2010-Hum dataset; bBold indicates the optimal value of SML2010-outTem dataset.
Back to article page