Bayesian learning of weakly structural Markov graph laws using sequential Monte Carlo methods
We present a sequential sampling methodology for weakly structural Markov laws, arising naturally in a Bayesian structure learning context for decomposable graphical models. As a key component of our suggested approach, we show that the problem of graph estimation, which in general lacks natural seq...
Saved in:
Published in | Electronic journal of statistics Vol. 13; no. 2; p. 2865 |
---|---|
Main Authors | , , |
Format | Journal Article |
Language | English |
Published |
01.01.2019
|
Subjects | |
Online Access | Get full text |
Cover
Loading…
Abstract | We present a sequential sampling methodology for weakly structural Markov laws, arising naturally in a Bayesian structure learning context for decomposable graphical models. As a key component of our suggested approach, we show that the problem of graph estimation, which in general lacks natural sequential interpretation, can be recast into a sequential setting by proposing a recursive Feynman-Kac model that generates a flow of junction tree distributions over a space of increasing dimensions. We focus on particle McMC methods to provide samples on this space, in particular on particle Gibbs (PG), as it allows for generating McMC chains with global moves on an underlying space of decomposable graphs. To further improve the PG mixing properties, we incorporate a systematic refreshment step implemented through direct sampling from a backward kernel. The theoretical properties of the algorithm are investigated, showing that the proposed refreshment step improves the performance in terms of asymptotic variance of the estimated distribution. The suggested sampling methodology is illustrated through a collection of numerical examples demonstrating high accuracy in Bayesian graph structure learning in both discrete and continuous graphical models. |
---|---|
AbstractList | We present a sequential sampling methodology for weakly structural Markov laws, arising naturally in a Bayesian structure learning context for decomposable graphical models. As a key component of our suggested approach, we show that the problem of graph estimation, which in general lacks natural sequential interpretation, can be recast into a sequential setting by proposing a recursive Feynman-Kac model that generates a flow of junction tree distributions over a space of increasing dimensions. We focus on particle McMC methods to provide samples on this space, in particular on particle Gibbs (PG), as it allows for generating McMC chains with global moves on an underlying space of decomposable graphs. To further improve the PG mixing properties, we incorporate a systematic refreshment step implemented through direct sampling from a backward kernel. The theoretical properties of the algorithm are investigated, showing that the proposed refreshment step improves the performance in terms of asymptotic variance of the estimated distribution. The suggested sampling methodology is illustrated through a collection of numerical examples demonstrating high accuracy in Bayesian graph structure learning in both discrete and continuous graphical models. |
Author | Olsson, Jimmy Pavlenko, Tatjana Rios, Felix L. |
Author_xml | – sequence: 1 givenname: Jimmy surname: Olsson fullname: Olsson, Jimmy – sequence: 2 givenname: Tatjana surname: Pavlenko fullname: Pavlenko, Tatjana – sequence: 3 givenname: Felix L. surname: Rios fullname: Rios, Felix L. |
BackLink | https://urn.kb.se/resolve?urn=urn:nbn:se:kth:diva-268610$$DView record from Swedish Publication Index |
BookMark | eNpN0DtPwzAUBWALFYm2sPALPCMFfJPYccZSyktFDDw2ZN04Thua2sVOqPrvIWqFmM4ZvnuHMyID66wh5BzYJcSQXkEezR5fgEt-RIaQJzzKeJwO_vUTMgrhkzEuYyGG5OMadybUaGlj0NvaLqir6NbgqtnR0PpOt53Hhj6hX7lvuvC4WdIGt4F2ocfBfHXGtnVPnG0NnaJvHF2bdunKcEqOK2yCOTvkmLzdzl6n99H8-e5hOplHOmG8jTKdZwxSWSV5CVCBKTJRlMgKXmAqdKmFZoWsQEKuZZXlWv8y0KlkhUYpWDIm0f5v2JpNV6iNr9fod8phrW7q94lyfqFW7VLFQgro_cXea-9C8Kb6uwCm-iEV5OowZPID-bVpog |
CitedBy_id | crossref_primary_10_1090_tpms_1198 crossref_primary_10_1007_s11222_022_10113_2 |
Cites_doi | 10.1007/s11222-014-9541-6 10.1093/biomet/86.3.615 10.3150/14-BEJ629 10.1214/11-BA630 10.1093/biomet/asx072 10.1214/14-AOS1209 10.1080/01621459.1994.10476894 10.1214/aos/1176349846 10.1214/08-AOS669 10.1007/s10479-007-0190-4 10.1198/jcgs.2009.07129 10.1007/0-387-28982-8 10.1093/biomet/86.4.785 10.1111/j.1467-9868.2009.00736.x 10.1214/aos/1176349260 10.1093/biomet/72.2.339 10.1093/biomet/ass052 10.1214/088342305000000304 10.1111/j.1467-9868.2006.00553.x |
ContentType | Journal Article |
DBID | AAYXX CITATION ADTPV AFDQA AOWAS D8T D8V ZZAVC |
DOI | 10.1214/19-EJS1585 |
DatabaseName | CrossRef SwePub SWEPUB Kungliga Tekniska Högskolan full text SwePub Articles SWEPUB Freely available online SWEPUB Kungliga Tekniska Högskolan SwePub Articles full text |
DatabaseTitle | CrossRef |
DatabaseTitleList | |
DeliveryMethod | fulltext_linktorsrc |
Discipline | Mathematics |
EISSN | 1935-7524 |
ExternalDocumentID | oai_DiVA_org_kth_268610 10_1214_19_EJS1585 |
GroupedDBID | 2WC 5VS AAYXX AENEX AFFOW ALMA_UNASSIGNED_HOLDINGS CITATION CS3 E3Z GR0 GROUPED_DOAJ J9A KQ8 M~E OK1 P2P RBV RPE TR2 ADTPV AFDQA AOWAS C1A D8T D8V ZZAVC |
ID | FETCH-LOGICAL-c305t-7c970148f39d11f1eb76bda0b5ba46cdc6c0b8f1819c8f79ccd111c480bca8603 |
ISSN | 1935-7524 |
IngestDate | Tue Apr 09 20:40:51 EDT 2024 Fri Aug 23 00:34:35 EDT 2024 |
IsDoiOpenAccess | true |
IsOpenAccess | true |
IsPeerReviewed | true |
IsScholarly | true |
Issue | 2 |
Language | English |
LinkModel | OpenURL |
MergedId | FETCHMERGED-LOGICAL-c305t-7c970148f39d11f1eb76bda0b5ba46cdc6c0b8f1819c8f79ccd111c480bca8603 |
OpenAccessLink | http://dx.doi.org/10.1214/19-EJS1585 |
ParticipantIDs | swepub_primary_oai_DiVA_org_kth_268610 crossref_primary_10_1214_19_EJS1585 |
PublicationCentury | 2000 |
PublicationDate | 2019-1-1 2019 |
PublicationDateYYYYMMDD | 2019-01-01 |
PublicationDate_xml | – month: 01 year: 2019 text: 2019-1-1 day: 01 |
PublicationDecade | 2010 |
PublicationTitle | Electronic journal of statistics |
PublicationYear | 2019 |
References | 22 11 12 23 13 24 14 15 16 17 18 19 1 2 3 4 5 6 7 8 9 20 10 21 |
References_xml | – ident: 23 doi: 10.1007/s11222-014-9541-6 – ident: 7 doi: 10.1093/biomet/86.3.615 – ident: 4 doi: 10.3150/14-BEJ629 – ident: 2 doi: 10.1214/11-BA630 – ident: 14 doi: 10.1093/biomet/asx072 – ident: 18 doi: 10.1214/14-AOS1209 – ident: 11 – ident: 17 doi: 10.1080/01621459.1994.10476894 – ident: 22 doi: 10.1214/aos/1176349846 – ident: 20 doi: 10.1214/08-AOS669 – ident: 10 – ident: 16 – ident: 19 doi: 10.1007/s10479-007-0190-4 – ident: 24 doi: 10.1198/jcgs.2009.07129 – ident: 3 doi: 10.1007/0-387-28982-8 – ident: 12 doi: 10.1093/biomet/86.4.785 – ident: 1 doi: 10.1111/j.1467-9868.2009.00736.x – ident: 5 doi: 10.1214/aos/1176349260 – ident: 9 doi: 10.1093/biomet/72.2.339 – ident: 13 doi: 10.1093/biomet/ass052 – ident: 15 doi: 10.1214/088342305000000304 – ident: 8 – ident: 6 doi: 10.1111/j.1467-9868.2006.00553.x – ident: 21 |
SSID | ssj0058266 |
Score | 2.1888044 |
Snippet | We present a sequential sampling methodology for weakly structural Markov laws, arising naturally in a Bayesian structure learning context for decomposable... |
SourceID | swepub crossref |
SourceType | Open Access Repository Aggregation Database |
StartPage | 2865 |
SubjectTerms | Decomposable graphical models Particle gibbs Sequential sampling Structure learning |
Title | Bayesian learning of weakly structural Markov graph laws using sequential Monte Carlo methods |
URI | https://urn.kb.se/resolve?urn=urn:nbn:se:kth:diva-268610 |
Volume | 13 |
hasFullText | 1 |
inHoldings | 1 |
isFullTextHit | |
isPrint | |
link | http://utb.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwnV3fi9QwEA5yvuiD-BPPXwQUX0rPJk3T5PHUO45FRXRP7kVKkia6bt2KW-88_3onSbvdxRVOX8oSQrvMfEwmw3zfIPSEOqUtNS5lZVmnTAueCupYSgtXKkNdRk1okH3Dj47Z5KQ4GbkngV3S6T3zayuv5H-8CmvgV8-S_QfPrl4KC_Ab_AtP8DA8L-Tj5-rcBhJkMxQ4IPU7s2renCdRGDaIang-TnuaBG3qpFFny-RHqBDENuou0HO9SJXv_mjafqj0cqNmPw7LWZOa8GSkqPM8lmqXPYNrMvs6ciXeqlM43OahKjtV3Re1WB0G72axz-_QNrOfySvrZzbU7Xotoo91MXBKL31ZRD70nt2yNkTbfA1VdD10ijg04o-YTgnz5QWZHkzekyIO-Nmikf1y9mG_ar9_qubd54pywT3x7jItZcGGm3Y8oAu4UfHYbBD_Xq9aC995Nn5lI0_ZUJENmcf0OrrWXxnwfvT_DXTJLm6iq69XervLW-jjgAQ8IAG3Dkck4BEJOCIBByRgjwQckIBHJOCABByQgHsk3EbHhwfTF0dpPzgjNRC-u7Q0svSVYpfLmhBHrC65rlWmC60YN7XhJtPCQXInjXClNAa2EcNEpo0SPMvvoJ1Fu7B3ETacSaWNrk3OmGREKe1oLWhNjCY6t7vo8WCm6lvUR6n8vRKMWRFZ9cbcRU-jBVd7_uKxexfdeB9d8QCMxbAHaAcsaR9CetjpR8HZvwHHe2_J |
link.rule.ids | 230,315,786,790,870,891,4043,27954,27955,27956 |
linkProvider | Directory of Open Access Journals |
openUrl | ctx_ver=Z39.88-2004&ctx_enc=info%3Aofi%2Fenc%3AUTF-8&rfr_id=info%3Asid%2Fsummon.serialssolutions.com&rft_val_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Ajournal&rft.genre=article&rft.atitle=Bayesian+learning+of+weakly+structural+Markov+graph+laws+using+sequential+Monte+Carlo+methods&rft.jtitle=Electronic+journal+of+statistics&rft.au=Olsson%2C+Jimmy&rft.au=Pavlenko%2C+Tatjana&rft.au=Rios%2C+Felix+Leopoldo&rft.date=2019&rft.issn=1935-7524&rft.eissn=1935-7524&rft.volume=13&rft.issue=2&rft.spage=2865&rft_id=info:doi/10.1214%2F19-EJS1585&rft.externalDocID=oai_DiVA_org_kth_268610 |
thumbnail_l | http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/lc.gif&issn=1935-7524&client=summon |
thumbnail_m | http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/mc.gif&issn=1935-7524&client=summon |
thumbnail_s | http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/sc.gif&issn=1935-7524&client=summon |