A short review on predictions for wind power generation – its limitation and future directions

Controlling and predicting power generation are essential elements for ef-ficient smart grid operations. In the case of wind energy, accurate prediction is even more important because humans have no control over its energy source. This study first reviews the line of prediction studies. Then, we generate point estimates using several machine learning methods and compare the performance and the pattern of the generated forecasts. We discuss the limitations of the point estimates and why probabilistic predictions are desirable. We suggest a few considerations for future studies.