Table 1 Relevant recent researches in weather prediction.

From: Deep learning approaches for predicting solar radiation and freshwater yield in modified pyramid solar still

References

Year

Models

Target feature

Forecast horizon

Time period

Location

Best evaluation parameters

37

2024

LSTM, GRU, BiLSTM

Temperature

Short-term

1901 to 2020

Saudi Arabia

RMSE = 0.0388 °C

MSE = 0.0656 °C

MAE = 0.0297 °C

R2 = 97.152%

38

2018

LSTM, GRU, BiLSTM

Rainfall

70 months

–

10,368 Geographic Locations around the globe

RMSE = 2.44

MAPE = 1.72

39

2022

LSTM, GRU, BiGRU, NBEATS

Temperature

Short-term

2015 to 2022

Bhubaneswar, India

MAE = 0.8097 °C

MSE = 1.1095 °C

RMSE = 1.1053 °C

R2 = 92.88%

40

2023

LSTM, GRU

Precipitation

Short-term

2005 to 2022

Jakarta

RMSE = 17.99

R2 =—0.07

41

2023

FFNN, LSTM, BiLSTM

Solar irradiance

Day-ahead

2013 to 2022

Seoul, South Korea

RMSE = 50.4 W/m2

NRMSE = 4.8 W/m2

MAE = 19.8 W/m2

NMAE = 1.9 W/m2

42

2023

CNN, MLP, LSTM, GRU, CNN-LSTM

Solar irradiance, Temperature

10 years

1984 to 2021

Zahedan, Iran

RMSE = 10.46 W/m2, 1.22° C

MAE = 8.15 W/m2, 0.95° C

R2 = 0.9722%, 0.9745%

NRMSE = 4.28 W/m2, 6.86° C

43

2023

CNN, GRU, LSTM, CNN-LSTM

Wind speed

6, 12, 60 months

1984 to 2021

Zabol, Iran

MAE = 0.483 m/s

R2 = 92.43%

RMSE = 0.722 m/s

COV = 9.58%

EC = 93.93%

44

2021

CNN-GRU-RPASM, CNN-GRU, CNN, RNN, GRU, LSTM

Temperature

1 to 24 h

2018 to 2019

China

MAE = 1.79 °C