Augmented Functional Time Series Representation and Forecasting with Gaussian Processes

Part of Advances in Neural Information Processing Systems 20 (NIPS 2007)

Bibtex Metadata Paper Supplemental

Authors

Nicolas Chapados, Yoshua Bengio

Abstract

We introduce a functional representation of time series which allows forecasts to be performed over an unspeciļ¬ed horizon with progressively-revealed informa- tion sets. By virtue of using Gaussian processes, a complete covariance matrix between forecasts at several time-steps is available. This information is put to use in an application to actively trade price spreads between commodity futures con- tracts. The approach delivers impressive out-of-sample risk-adjusted returns after transaction costs on a portfolio of 30 spreads.