Wind Energy capacity and generation forecasting: Machine Learning-based projections for the International Markets to 2030

As a clean and renewable energy source, wind energy helps combat emissions and climate change. This study presents a comprehensive machine learning approach to forecast the worldwide wind energy capacity and generation through 2030 using the gradient boosting decision tree (GBDT) methodologies. An analysis of 20 leading wind-energy-producing countries reveals a strong linear relationship between the installed capacity and the actual generation ( ), which indicates that the proposed models capture this aggregate capacity-generation link robustly at the cross-country level. The results demonstrate that China will maintain dominance with approximately 60% of the worldwide installed capacity (under the accelerated scenario) and about 51-52% of the generation output by 2030, while the United States and Germany come second and third respectively for the capacity and the generation outputs, reflecting the superior capacity factors in the developed markets. The three organized scenarios (the conservative, the reference, and the accelerated) that are performed in this study together with the three capacity-factor assumptions (the pessimistic, the central, and the optimistic) disclose that the installed capacity trajectories are the primary driver of the generation-share distributions.