Rao-Blackwellized Particle MCMC for Parameter Estimation in Spatio-Temporal Gaussian Processes

Roland Hostettler; Simo Särkkä; Simon J. Godsill

doi:10.1109/MLSP.2017.8168171

Rao-Blackwellized Particle MCMC for Parameter Estimation in Spatio-Temporal Gaussian Processes

Roland Hostettler, Simo Särkkä, Simon J. Godsill

University of Cambridge

Tutkimustuotos: Artikkeli kirjassa/konferenssijulkaisussa › Conference article in proceedings › Scientific › vertaisarvioitu

3 Sitaatiot (Scopus)

211 Lataukset (Pure)

Abstrakti

In this paper, we consider parameter estimation in latent, spatiotemporal Gaussian processes using particle Markov chain Monte Carlo methods. In particular, we use spectral decomposition of the covariance function to obtain a high-dimensional state-space representation of the Gaussian processes, which is assumed to be observed through a nonlinear non-Gaussian likelihood. We develop a Rao-Blackwellized particle Gibbs sampler to sample the state trajectory and show how to sample the hyperparameters and possible parameters in the likelihood. The proposed method is evaluated on a spatio-temporal population model and the predictive performance is evaluated using leave-one-out cross-validation.

Alkuperäiskieli	Englanti
Otsikko	Proceedings of the 27th IEEE International Workshop on Machine Learning for Signal Processing, MLSP 2017
Kustantaja	IEEE
Sivumäärä	6
ISBN (elektroninen)	9781509063413
DOI - pysyväislinkit	https://doi.org/10.1109/MLSP.2017.8168171
Tila	Julkaistu - 2017
OKM-julkaisutyyppi	A4 Artikkeli konferenssijulkaisussa
Tapahtuma	IEEE International Workshop on Machine Learning for Signal Processing - Tokyo, Japani Kesto: 25 syysk. 2017 → 28 syysk. 2017 Konferenssinumero: 27 http://mlsp2017.conwiz.dk/home.htm

Julkaisusarja

Nimi	IEEE International Workshop on Machine Learning for Signal Processing
Kustantaja	IEEE
ISSN (painettu)	2161-0363
ISSN (elektroninen)	2161-0371

Workshop

Workshop	IEEE International Workshop on Machine Learning for Signal Processing
Lyhennettä	MLSP
Maa/Alue	Japani
Kaupunki	Tokyo
Ajanjakso	25/09/2017 → 28/09/2017
www-osoite	http://mlsp2017.conwiz.dk/home.htm

Pääsy asiakirjaan

10.1109/MLSP.2017.8168171

2017-mlspAccepted author manuscript, 354 KB

Siteeraa tätä

Hostettler, R., Särkkä, S., & Godsill, S. J. (2017). Rao-Blackwellized Particle MCMC for Parameter Estimation in Spatio-Temporal Gaussian Processes. teoksessa Proceedings of the 27th IEEE International Workshop on Machine Learning for Signal Processing, MLSP 2017 (IEEE International Workshop on Machine Learning for Signal Processing). IEEE. https://doi.org/10.1109/MLSP.2017.8168171

@inproceedings{213b5e3c8b6a4c24b1d165d0affa1f37,

title = "Rao-Blackwellized Particle MCMC for Parameter Estimation in Spatio-Temporal Gaussian Processes",

abstract = "In this paper, we consider parameter estimation in latent, spatiotemporal Gaussian processes using particle Markov chain Monte Carlo methods. In particular, we use spectral decomposition of the covariance function to obtain a high-dimensional state-space representation of the Gaussian processes, which is assumed to be observed through a nonlinear non-Gaussian likelihood. We develop a Rao-Blackwellized particle Gibbs sampler to sample the state trajectory and show how to sample the hyperparameters and possible parameters in the likelihood. The proposed method is evaluated on a spatio-temporal population model and the predictive performance is evaluated using leave-one-out cross-validation.",

author = "Roland Hostettler and Simo S{\"a}rkk{\"a} and Godsill, {Simon J.}",

year = "2017",

doi = "10.1109/MLSP.2017.8168171",

language = "English",

series = "IEEE International Workshop on Machine Learning for Signal Processing",

publisher = "IEEE",

booktitle = "Proceedings of the 27th IEEE International Workshop on Machine Learning for Signal Processing, MLSP 2017",

address = "United States",

note = "IEEE International Workshop on Machine Learning for Signal Processing , MLSP ; Conference date: 25-09-2017 Through 28-09-2017",

url = "http://mlsp2017.conwiz.dk/home.htm",

}

Hostettler, R, Särkkä, S & Godsill, SJ 2017, Rao-Blackwellized Particle MCMC for Parameter Estimation in Spatio-Temporal Gaussian Processes. julkaisussa Proceedings of the 27th IEEE International Workshop on Machine Learning for Signal Processing, MLSP 2017 . IEEE International Workshop on Machine Learning for Signal Processing, IEEE, IEEE International Workshop on Machine Learning for Signal Processing , Tokyo, Japani, 25/09/2017. https://doi.org/10.1109/MLSP.2017.8168171

Rao-Blackwellized Particle MCMC for Parameter Estimation in Spatio-Temporal Gaussian Processes. / Hostettler, Roland; Särkkä, Simo; Godsill, Simon J.
Proceedings of the 27th IEEE International Workshop on Machine Learning for Signal Processing, MLSP 2017 . IEEE, 2017. (IEEE International Workshop on Machine Learning for Signal Processing).

Tutkimustuotos: Artikkeli kirjassa/konferenssijulkaisussa › Conference article in proceedings › Scientific › vertaisarvioitu

TY - GEN

T1 - Rao-Blackwellized Particle MCMC for Parameter Estimation in Spatio-Temporal Gaussian Processes

AU - Hostettler, Roland

AU - Särkkä, Simo

AU - Godsill, Simon J.

N1 - Conference code: 27

PY - 2017

Y1 - 2017

N2 - In this paper, we consider parameter estimation in latent, spatiotemporal Gaussian processes using particle Markov chain Monte Carlo methods. In particular, we use spectral decomposition of the covariance function to obtain a high-dimensional state-space representation of the Gaussian processes, which is assumed to be observed through a nonlinear non-Gaussian likelihood. We develop a Rao-Blackwellized particle Gibbs sampler to sample the state trajectory and show how to sample the hyperparameters and possible parameters in the likelihood. The proposed method is evaluated on a spatio-temporal population model and the predictive performance is evaluated using leave-one-out cross-validation.

AB - In this paper, we consider parameter estimation in latent, spatiotemporal Gaussian processes using particle Markov chain Monte Carlo methods. In particular, we use spectral decomposition of the covariance function to obtain a high-dimensional state-space representation of the Gaussian processes, which is assumed to be observed through a nonlinear non-Gaussian likelihood. We develop a Rao-Blackwellized particle Gibbs sampler to sample the state trajectory and show how to sample the hyperparameters and possible parameters in the likelihood. The proposed method is evaluated on a spatio-temporal population model and the predictive performance is evaluated using leave-one-out cross-validation.

U2 - 10.1109/MLSP.2017.8168171

DO - 10.1109/MLSP.2017.8168171

M3 - Conference article in proceedings

T3 - IEEE International Workshop on Machine Learning for Signal Processing

BT - Proceedings of the 27th IEEE International Workshop on Machine Learning for Signal Processing, MLSP 2017

PB - IEEE

T2 - IEEE International Workshop on Machine Learning for Signal Processing

Y2 - 25 September 2017 through 28 September 2017

ER -

Rao-Blackwellized Particle MCMC for Parameter Estimation in Spatio-Temporal Gaussian Processes

Abstrakti

Julkaisusarja

Workshop

Pääsy asiakirjaan

Sormenjälki

Siteeraa tätä