Show simple item record

dc.contributor.authorLee, Anthony
dc.contributor.authorSingh, Sumeetpal S.
dc.contributor.authorVihola, Matti
dc.date.accessioned2020-10-01T04:03:04Z
dc.date.available2020-10-01T04:03:04Z
dc.date.issued2020
dc.identifier.citationLee, A., Singh, S. S., & Vihola, M. (2020). Coupled conditional backward sampling particle filter. <i>Annals of Statistics</i>, <i>48</i>(5), 3066-3089. <a href="https://doi.org/10.1214/19-AOS1922" target="_blank">https://doi.org/10.1214/19-AOS1922</a>
dc.identifier.otherCONVID_42303180
dc.identifier.urihttps://jyx.jyu.fi/handle/123456789/71949
dc.description.abstractThe conditional particle filter (CPF) is a promising algorithm for general hidden Markov model smoothing. Empirical evidence suggests that the variant of CPF with backward sampling (CBPF) performs well even with long time series. Previous theoretical results have not been able to demonstrate the improvement brought by backward sampling, whereas we provide rates showing that CBPF can remain effective with a fixed number of particles independent of the time horizon. Our result is based on analysis of a new coupling of two CBPFs, the coupled conditional backward sampling particle filter (CCBPF). We show that CCBPF has good stability properties in the sense that with fixed number of particles, the coupling time in terms of iterations increases only linearly with respect to the time horizon under a general (strong mixing) condition. The CCBPF is useful not only as a theoretical tool, but also as a practical method that allows for unbiased estimation of smoothing expectations, following the recent developments by Jacob, Lindsten and Schon (2020). Unbiased estimation has many advantages, such as enabling the construction of asymptotically exact confidence intervals and straightforward parallelisation.en
dc.format.mimetypeapplication/pdf
dc.languageeng
dc.language.isoeng
dc.publisherInstitute of Mathematical Statistics
dc.relation.ispartofseriesAnnals of Statistics
dc.rightsIn Copyright
dc.subject.otherbackward sampling
dc.subject.otherconvergence rate
dc.subject.othercoupling
dc.subject.otherconditional particle filter
dc.subject.otherunbiased
dc.titleCoupled conditional backward sampling particle filter
dc.typearticle
dc.identifier.urnURN:NBN:fi:jyu-202010016029
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.description.reviewstatuspeerReviewed
dc.format.pagerange3066-3089
dc.relation.issn0090-5364
dc.relation.numberinseries5
dc.relation.volume48
dc.type.versionpublishedVersion
dc.rights.copyright© Institute of Mathematical Statistics, 2020
dc.rights.accesslevelopenAccessfi
dc.relation.grantnumber315619
dc.relation.grantnumber284513
dc.relation.grantnumber274740
dc.relation.grantnumber312605
dc.subject.ysonumeerinen analyysi
dc.subject.ysoMarkovin ketjut
dc.subject.ysostokastiset prosessit
dc.subject.ysoMonte Carlo -menetelmät
dc.format.contentfulltext
jyx.subject.urihttp://www.yso.fi/onto/yso/p15833
jyx.subject.urihttp://www.yso.fi/onto/yso/p13075
jyx.subject.urihttp://www.yso.fi/onto/yso/p11400
jyx.subject.urihttp://www.yso.fi/onto/yso/p6361
dc.rights.urlhttp://rightsstatements.org/page/InC/1.0/?language=en
dc.relation.doi10.1214/19-AOS1922
dc.relation.funderSuomen Akatemiafi
dc.relation.funderSuomen Akatemiafi
dc.relation.funderSuomen Akatemiafi
dc.relation.funderSuomen Akatemiafi
dc.relation.funderAcademy of Finlanden
dc.relation.funderAcademy of Finlanden
dc.relation.funderAcademy of Finlanden
dc.relation.funderAcademy of Finlanden
jyx.fundingprogramAkatemiahanke, SAfi
jyx.fundingprogramAkatemiatutkijan tutkimuskulut, SAfi
jyx.fundingprogramAkatemiatutkijan tehtävä, SAfi
jyx.fundingprogramAkatemiatutkijan tutkimuskulut, SAfi
jyx.fundingprogramAcademy Project, AoFen
jyx.fundingprogramResearch costs of Academy Research Fellow, AoFen
jyx.fundingprogramResearch post as Academy Research Fellow, AoFen
jyx.fundingprogramResearch costs of Academy Research Fellow, AoFen
jyx.fundinginformationThis work wassupported by EPSRC grant numbers EP/K032208/1, EP/R014604/1 and EP/R034710/1, and by the Alan Turing Institute under the EPSRC grant EP/N510129/1. MV was supported byAcademy of Finland grants 274740, 284513, 312605 and 315619. The authors wish to ac-knowledge CSC, IT Center for Science, Finland, for computational resources.


Files in this item

Thumbnail

This item appears in the following Collection(s)

Show simple item record

In Copyright
Except where otherwise noted, this item's license is described as In Copyright