Machine Learning as a Downscaling Approach for Prediction of Wind Characteristics under Future Climate Change Scenarios

Table 1

Developed models in the present study and their description.

Model type	Description
A (Annual)	One DT model was developed using the Train data, validated by the Test data, and adopted for prediction in the forecast section.
S (Seasonal)	The entire data were categorized into seasons, and four decision tree (DT) models were generated alongside together. Each model is developed and validated with the data of the relevant season and used for forecasting in the season for the projection period. The names of these models based on the seasons are as follows: Swinter: winter’s model, Sspring: spring’s model, Ssummer: summer’s model, and Sfall: fall’s model.
M (Monthly)	The entire data were divided into months, and 12 DT models were extended beside together. Each model is developed and verified with the data of the relevant month and used for forecasting in the relevant month for the projection period. The names of these models based on the months are MJan to MDec models for January to December, respectively.