Posters
See the list of poster presenters at the Technology Workshop 2025 – and check out their work!
For more details on each poster, click on the poster titles to read the abstract.
PO022: Tree-based Spatio-temporal Feature Learning for Short-term Wind Power Forecasting Tasks
Selahattin Seha Cirit, Meteorology Data Scientist, ALGOPOLY & BOGAZICI UNIVERSITY
Abstract
Accurate wind power forecasting plays an important role in the optimization of renewable energy grid management to ensure a stable energy supply. Numerical Weather Predictions (NWP) serve as the main input for the forecasting models; however, NWP data poses significant challenges in modeling due to the high-dimensionality arising from the numerous locations (i.e., latitude, longitude), levels (i.e., altitude), and variables (i.e., temperature, wind speed). This study proposes a novel spatio-temporal feature learning algorithm for efficient and accurate forecasting by embedding the predictions from multiple NWP models. Our proposal integrates supervised and unsupervised embeddings into tree-based learning to model spatial and temporal dependencies between the outputs of NWP data. The proposed methodology was tested on a short-term wind power forecasting task for wind farms in the Aegean Region of Türkiye. Our experiments show that the proposed feature learning strategy provides competitive results in terms of accuracy with a significantly reduced number of features.
No recording available for this poster.
