Improvement of short-term forecast for wind speed

Abstract

This paper presents a new type of linear regression model called sparse linear regression (SLR) model for short-term wind speed forecasting. Modifications are applied to the SLR model and some other variant models are proposed. Experiments are carried out on real wind farm history recording data. Results show SLR model and its variants can improve the accuracy of the short-term forecasting result compared with linear regression model.