Calibration of Multifactor Models in Electricity Markets

Spot prices of electricity and other energy commodities are often modeled by multifactor stochastic processes. This poses a problem of estimating models' parameters based on historical data, i.e. calibrating them to markets. Here we show how a traditional tool of Kalman Filters can be successfully applied to do this task. We study two mean-reverting log-spot price models and the Pilipovic model using correspondingly Kalman Filter and the extended Kalman Filter. The results of applying this method to market data from several power exchanges are discussed