Show simple item record

dc.contributor.authorKarppinen, Santeri
dc.contributor.authorVihola, Matti
dc.date.accessioned2021-03-05T08:47:44Z
dc.date.available2021-03-05T08:47:44Z
dc.date.issued2021
dc.identifier.citationKarppinen, S., & Vihola, M. (2021). Conditional particle filters with diffuse initial distributions. <i>Statistics and Computing</i>, <i>31</i>(3), Article 24. <a href="https://doi.org/10.1007/s11222-020-09975-1" target="_blank">https://doi.org/10.1007/s11222-020-09975-1</a>
dc.identifier.otherCONVID_51778162
dc.identifier.urihttps://jyx.jyu.fi/handle/123456789/74498
dc.description.abstractConditional particle filters (CPFs) are powerful smoothing algorithms for general nonlinear/non-Gaussian hidden Markov models. However, CPFs can be inefficient or difficult to apply with diffuse initial distributions, which are common in statistical applications. We propose a simple but generally applicable auxiliary variable method, which can be used together with the CPF in order to perform efficient inference with diffuse initial distributions. The method only requires simulatable Markov transitions that are reversible with respect to the initial distribution, which can be improper. We focus in particular on random walk type transitions which are reversible with respect to a uniform initial distribution (on some domain), and autoregressive kernels for Gaussian initial distributions. We propose to use online adaptations within the methods. In the case of random walk transition, our adaptations use the estimated covariance and acceptance rate adaptation, and we detail their theoretical validity. We tested our methods with a linear Gaussian random walk model, a stochastic volatility model, and a stochastic epidemic compartment model with time-varying transmission rate. The experimental findings demonstrate that our method works reliably with little user specification and can be substantially better mixing than a direct particle Gibbs algorithm that treats initial states as parameters.en
dc.format.mimetypeapplication/pdf
dc.languageeng
dc.language.isoeng
dc.publisherSpringer
dc.relation.ispartofseriesStatistics and Computing
dc.rightsCC BY 4.0
dc.subject.otherAdaptive Markov chain Monte Carlo
dc.subject.otherbayesian inference
dc.subject.othercompartment model
dc.subject.otherconditional particle filter
dc.subject.otherdiffuse initialisation
dc.subject.otherHidden Markov model
dc.subject.othersmoothing
dc.subject.otherstate space model
dc.titleConditional particle filters with diffuse initial distributions
dc.typearticle
dc.identifier.urnURN:NBN:fi:jyu-202103051858
dc.contributor.laitosMatematiikan ja tilastotieteen laitosfi
dc.contributor.laitosDepartment of Mathematics and Statisticsen
dc.contributor.oppiaineTilastotiedefi
dc.contributor.oppiaineStatisticsen
dc.type.urihttp://purl.org/eprint/type/JournalArticle
dc.relation.isbn0960-3174
dc.description.reviewstatuspeerReviewed
dc.relation.issn0960-3174
dc.relation.numberinseries3
dc.relation.volume31
dc.type.versionpublishedVersion
dc.rights.copyright© The Author(s) 2021
dc.rights.accesslevelopenAccessfi
dc.relation.grantnumber315619
dc.subject.ysobayesilainen menetelmä
dc.subject.ysoMarkovin ketjut
dc.subject.ysomatemaattiset menetelmät
dc.subject.ysotilastotiede
dc.format.contentfulltext
jyx.subject.urihttp://www.yso.fi/onto/yso/p17803
jyx.subject.urihttp://www.yso.fi/onto/yso/p13075
jyx.subject.urihttp://www.yso.fi/onto/yso/p1880
jyx.subject.urihttp://www.yso.fi/onto/yso/p3591
dc.rights.urlhttps://creativecommons.org/licenses/by/4.0/
dc.relation.doi10.1007/s11222-020-09975-1
dc.relation.funderSuomen Akatemiafi
dc.relation.funderAcademy of Finlanden
jyx.fundingprogramAkatemiahanke, SAfi
jyx.fundingprogramAcademy Project, AoFen
jyx.fundinginformationThis work was supported by Academy of Finland Grant 315619.


Files in this item

Thumbnail

This item appears in the following Collection(s)

Show simple item record

CC BY 4.0
Except where otherwise noted, this item's license is described as CC BY 4.0