Data-driven modeling for long-term electricity price forecasting

Estimating the financial viability of renewable energy investments requires availability long-term, finely-resolved electricity prices over investment lifespan. This entails, however, two major challenges: (i) combination extensive time horizons and fine resolutions, (ii) prediction out-of-sample in future market scenarios, or shifts pricing regime, that were not observed past. paper tackles such challenges by proposing a data-driven model for long-term is based on Fourier analysis. The price decomposed into components leading to its base evolution, which are described through amplitudes main frequencies series, high volatility, residual frequencies. former predicted via regression uses as input annual values relevant quantities, generation, demands. proposed method shows capable predicting most dynamics price; generalization capturing mechanisms previously unseen markets. These findings support relevance validity data-driven, finely-resolved, predictions highlight potential hybrid market-based models.