Bayesian calibration of the Schwartz-Smith Model adapted to the energy market
2014 (English)In: 2014 IEEE WORKSHOP ON STATISTICAL SIGNAL PROCESSING (SSP), IEEE , 2014, 508-511 p.Conference paper (Refereed)
We consider an application of Bayesian signal processing to the energy trading problem. In particular, we address the problem of calibrating the Schwartz-Smith Model using the observed electricity futures prices traded on the markets. As compared with the other financial markets, basic electricity derivatives such as futures are more complicated, as these products are based not on the spot prices themselves but on the arithmetic averages of the spot prices during the delivery period. As a result, the (log) futures prices are no longer affine function of the model factors and as such, an approach based on Kalman filtering, to estimate the latent model factors and the parameters seems meaningless. Here, we envisage a Bayesian approach using the particle marginal Metropolis Hastings (PMMH) algorithm for this challenging estimation task. We demonstrate the efficacy of our approach on simulated data.
Place, publisher, year, edition, pages
IEEE , 2014. 508-511 p.
PMMH, PMCMC, SMC, Schwartz-Smith model, financial signal processing
IdentifiersURN: urn:nbn:se:liu:diva-107392DOI: 10.1109/SSP.2014.6884687ISI: 000361019700128OAI: oai:DiVA.org:liu-107392DiVA: diva2:723825
IEEE statistical signal processing workshop (SSP 2014), Gold Coast, Australia, 29 June – 2 July 2014
FunderSwedish Research Council, CADICSSwedish Foundation for Strategic Research , COOPLOC