SEMINARS
|
|
Spring 2008
STATISTICS
COLLOQUIUM
Wednesday, April 16, 2008
3:30-4:00—Refreshments, Yost 327
4:00-5:00—Talk,
Yost 101
Rafal Weron, PhD
Institute of Mathematics and Computer Science
Wroclaw University of Technology
Modeling and forecasting electricity prices
In the first years after the emergence of deregulated power markets it became apparent that for the valuation of electricity contracts we cannot simply rely on models developed for financial or other commodity markets. However, before adequate models can be put forward the unique characteristics of electricity (spot) prices have to be thoroughly analyzed. In particular, the extreme volatility and price spikes which lead to heavy-tailed distributions of returns. In this paper we first analyze the stylized facts of electricity prices, then discuss adequate modeling approaches. In particular, we test whether jump-diffusion and regime-switching models with heavy-tailed components are better suited for modeling the electricity price process characteristics than classical models based on the Gaussian distribution. We also evaluate a collection of time series models in terms of their day-ahead forecasting performance and see whether non-linear or non-parametric approaches lead to better forecasts than classical ARMA-type models.
|
