Table 14 Hyperparameter-optimized forecasting performance of MSAN on wind dataset.

From: Optimizing solar and wind forecasting with iHow optimization algorithm and multi-scale attention networks

Model

MSE

RMSE

MAE

MBE

\(\varvec{r}\)

\(\varvec{R^2}\)

RRMSE

NSE

WI

iHOW + MSAN

1.11E−06

3.15E−07

7.34E−06

1.39E−06

0.9753

0.9815

4.58

0.9581

0.9457

HHO + MSAN

9.44E−05

9.43E−06

1.24E−05

1.63E−05

0.9589

0.9647

7.59

0.9451

0.9339

GWO + MSAN

1.06E−04

1.05E−05

1.27E−05

1.63E−05

0.9573

0.9628

8.54

0.9414

0.9265

PSO + MSAN

1.14E−04

1.16E−05

1.30E−05

1.64E−05

0.9557

0.9612

9.45

0.9395

0.9239

WAO + MSAN

1.14E−04

1.48E−05

1.39E−05

1.73E−05

0.9435

0.9604

10.17

0.9339

0.9219

BBO + MSAN

1.15E−04

1.75E−05

1.47E−05

1.75E−05

0.9426

0.9582

10.72

0.8799

0.9187

MVO + MSAN

1.18E−04

1.99E−05

1.55E−05

1.82E−05

0.9418

0.9540

11.27

0.9207

0.9203

SFS + MSAN

1.21E−04

2.24E−05

1.62E−05

1.88E−05

0.9395

0.9499

12.79

0.9177

0.9151

SAO + MSAN

1.31E−04

2.32E−05

1.64E−05

1.95E−05

0.9382

0.9488

13.45

0.9151

0.9137

JAYA + MSAN

1.55E−04

4.36E−05

1.70E−05

2.04E−05

0.9346

0.9476

13.85

0.9131

0.9127

Back to article page