Empirical Inference Conference Paper 2011

A general linear non-Gaussian state-space model: Identifiability, identification, and applications

State-space modeling provides a powerful tool for system identification and prediction. In linear state-space models the data are usually assumed to be Gaussian and the models have certain structural constraints such that they are identifiable. In this paper we propose a non-Gaussian state-space model which does not have such constraints. We prove that this model is fully identifiable. We then propose an efficient two-step method for parameter estimation: one first extracts the subspace of the latent processes based on the temporal information of the data, and then performs multichannel blind deconvolution, making use of both the temporal information and non-Gaussianity. We conduct a series of simulations to illustrate the performance of the proposed method. Finally, we apply the proposed model and parameter estimation method on real data, including major world stock indices and magnetoencephalography (MEG) recordings. Experimental results are encouraging and show the practical usefulness of the proposed model and method.

Author(s): Zhang, K. and Hyvärinen, A.
Book Title: JMLR Workshop and Conference Proceedings Volume 20
Pages: 113-128
Year: 2011
Month: November
Day: 0
Editors: Hsu, C.-N. , W.S. Lee
Publisher: MIT Press
Bibtex Type: Conference Paper (inproceedings)
Address: Cambridge, MA, USA
Event Name: 3rd Asian Conference on Machine Learning (ACML 2011)
Event Place: Taoyuan, Taiwan
Digital: 0
Electronic Archiving: grant_archive
Links:

BibTex

@inproceedings{ZhangH2011,
  title = {A general linear non-Gaussian state-space model: Identifiability, identification, and applications},
  booktitle = {JMLR Workshop and Conference Proceedings Volume 20},
  abstract = {State-space modeling provides a powerful tool for system identification and prediction. In linear state-space models the data are usually assumed to be Gaussian and the models have certain structural constraints such that they are identifiable. In this paper we propose a non-Gaussian state-space model which does not have such constraints. We prove that this model is fully identifiable. We then propose an efficient two-step method for parameter estimation: one first extracts the subspace of the latent processes based on the temporal information of the data, and then performs multichannel blind deconvolution, making use of both the temporal information and non-Gaussianity. We conduct a series of simulations to illustrate the performance of the proposed method. Finally, we apply the proposed model and parameter estimation method on real data, including major world stock indices and magnetoencephalography (MEG) recordings. Experimental results are encouraging and show the practical usefulness of the proposed model and method.},
  pages = {113-128},
  editors = {Hsu, C.-N. , W.S. Lee },
  publisher = {MIT Press},
  address = {Cambridge, MA, USA},
  month = nov,
  year = {2011},
  slug = {zhangh2011},
  author = {Zhang, K. and Hyv{\"a}rinen, A.},
  month_numeric = {11}
}