Uśrednianie i ocena dokładności w krótkoterminowym prognozowaniu cen energii elektrycznej  
[Expert aggregation and metrics for short-term electricity price forecasting]

David Obst (Department of Statistics & Mathematical Optimization, EDF R&D, Palaiseau, FRA)


Accurate electricity price forecasting (EPF) is crucial for grid management, optimal 
piloting of storage assets and computation of efficient trading strategies. However, 
predictions obtained from different models might have individual strengths and weaknesses, 
and varying performance over time. In this work, we present how expert aggregation allows 
to enhance forecasting accuracy over individual predictions for intraday and day-ahead 
(spot) prices. We present two types of aggregation methods: sequential aggregation and 
expert stacking. We will also discuss the relevance of different metrics to assess the 
quality of price forecasts from an operational point of view.