Table 2 MAE results of different models for WSP

From: A machine learning model for hub-height short-term wind speed prediction

Models

0 h

[m s−1]

1 h

[m s−1]

2 h

[m s−1]

3 h

[m s−1]

6 h

[m s−1]

9 h

[m s−1]

12 h

[m s−1]

18 h

[m s−1]

21 h

[m s−1]

23 h

[m s−1]

CNN

0.49

0.77

0.98

1.16

1.51

1.74

2.09

2.44

2.57

2.72

ConvLSTM

0.44

0.73

0.93

1.08

1.34

1.48

1.65

1.96

2.04

2.06

E3DLSTM

0.49

0.70

0.86

0.99

1.32

1.61

1.86

2.21

2.29

2.37

SimVP

0.35

0.56

0.70

0.81

1.05

1.24

1.45

1.76

1.86

1.94

TAU

0.37

0.56

0.69

0.80

1.02

1.21

1.42

1.71

1.82

1.91

Phydnet

2.00

2.10

2.19

2.27

2.44

2.45

2.44

2.69

2.68

2.61

STDGN

0.33

0.55

0.72

0.84

1.11

1.32

1.52

1.83

1.92

2.00

WPN

0.36

0.55

0.69

0.78

0.99

1.21

1.41

1.72

1.88

1.97

ECMWF

0.90

-

-

0.83

0.82

0.87

0.98

0.91

0.96

-

MFWPN

0.32

0.51

0.64

0.73

0.94

1.13

1.33

1.62

1.78

1.88

  1. 0 h, 1 h, and up to 23 h represent the hours from 00:00 to 23:00, with 0 h being the first hour predicted based on the input historical data.