AUEB Among the Top 10 Finalists at the Econometric Game 2026 | Open Seminar on the Finalist Paper: Multivariate Electricity Price Forecasting | Monday, 15 June | 15:00–16:30 | Troias Amphitheater
The Athens University of Economics and Business has once again achieved international distinction, advancing to the Top 10 Finalist Institutions at the 27th Econometric Game, hosted by the University of Amsterdam and widely regarded as the World Championship of Econometrics.
For the second consecutive year, AUEB secured a place among the competition’s top finalists, competing against teams from 30 leading universities worldwide such as Harvard, Oxford, National University Singapore etc.
Our PhD in Economics students, George Malanos (Team Leader) and Christina Logotheti, and Master’s students Vasileia Argyrou and George Skolarikis from the MSc in Business Economics with Analytics tackled a highly relevant global challenge in energy economics:
Multivariate electricity price forecasting using robust analytical models and dynamic, real-world data, including changing weather patterns, renewable energy production, and cross-border electricity flows.
The AUEB finalist paper will be presented at an open seminar for the university's academic community on Monday, 15 June (15:00 - 16:30) at the Troias Amphitheater. Those wishing to attend are kindly requested to complete the registration form available at: https://forms.gle/hEEyFyqvn6TA1nAc6
Methodological Innovation
To address the complexity of the problem, the team implemented a sophisticated multi-layered forecasting strategy combining econometric and machine-learning techniques:
- Core Modeling: A HAR/UMIDAS approach tailored specifically for mixed-frequency data.
- Predictions & Refinement: Tree-based Machine Learning methods for out-of-sample predictions, utilizing aggressive Recursive Feature Elimination for optimal parsimony.
- Explainability: Shapley Additive Values (SHAP) were used to translate complex models into concrete, actionable policy recommendations.
- Advanced Forecasting: STL Decomposition utilizing Harmonic Regressions, Seasonal Naïves, and Random Walks with drift to move past naive univariate approaches for unobservable out-of-sample variables.
This achievement reflects the strength of AUEB’s research-driven education and its commitment to developing analytical talent capable of addressing complex global challenges. We are proud to see our students applying cutting-edge econometric and data-analytics methods to real-world problems and representing AUEB with distinction on the international stage.
