Spatiotemporal wind forecasting by learning a hierarchically sparse inverse covariance matrix using wind directions

Given the advances in online data acquisition systems, statistical learning models are increasingly used to forecast wind speed. In electricity markets, farm production forecasts needed for day-ahead, intra-day, and real-time markets. this work, we use a spatiotemporal model that leverages dynamics Using priori knowledge of direction, propose maximum likelihood estimate inverse covariance matrix regularized with hierarchical sparsity-inducing penalty. The resulting not only exhibits benefits sparse estimator, but also enables meaningful structures by considering direction. A proximal method is solve underlying optimization problem. proposed methodology six-hour-ahead speeds 20-minute time intervals case study Texas. We compare our number other methods. Prediction performance measures Diebold–Mariano test show potential method, specifically when reasonably accurate estimates directions available.