English Translation: Data Splitting and Testing for Model Training

In order to prevent potential temporal dependencies, the data from January 2017 to December 2020 were used as the training and verification sets for model training. 20% of the data was randomly selected as the verification set, while the remaining 80% was used as the training set. We selected the data from the entire 21-year period as the testing set, not only because we needed to evaluate the forecasting performance of the completed year, including daily, monthly, and seasonal changes, but also because some results were not presented in the article in their entirety.