Endre søk
RefereraExporteraLink to record
Permanent link

Direct link
Referera
Referensformat
  • apa
  • harvard1
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • Annet format
Fler format
Språk
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Annet språk
Fler språk
Utmatningsformat
  • html
  • text
  • asciidoc
  • rtf
Identification of Hidden Markov Models Using Spectral Learning with Likelihood Maximization
KTH, Skolan för elektro- och systemteknik (EES), Reglerteknik.
KTH, Skolan för elektro- och systemteknik (EES), Reglerteknik.ORCID-id: 0000-0003-0355-2663
KTH, Skolan för elektro- och systemteknik (EES), Reglerteknik.
2017 (engelsk)Inngår i: 2017 IEEE 56th Annual Conference on Decision and Control, CDC 2017, Institute of Electrical and Electronics Engineers (IEEE), 2017, s. 5859-5864Konferansepaper, Publicerat paper (Fagfellevurdert)
Abstract [en]

In this paper, we consider identifying a hidden Markov model (HMM) with the purpose of computing estimates of joint and conditional (posterior) probabilities over observation sequences. The classical maximum likelihood estimation algorithm (via the Baum-Welch/expectation-maximization algorithm), has recently been challenged by methods of moments. Such methods employ low-order moments to provide parameter estimates and have several benefits, including consistency and low computational cost. This paper aims to reduce the gap in statistical efficiency that results from restricting to only low-order moments in the training data. In particular, we propose a two-step procedure that combines spectral learning with a single Newton-like iteration for maximum likelihood estimation. We demonstrate an improved statistical performance using the proposed algorithm in numerical simulations.

sted, utgiver, år, opplag, sider
Institute of Electrical and Electronics Engineers (IEEE), 2017. s. 5859-5864
Serie
IEEE Conference on Decision and Control, ISSN 0743-1546
HSV kategori
Identifikatorer
URN: urn:nbn:se:kth:diva-223860DOI: 10.1109/CDC.2017.8264545ISI: 000424696905103Scopus ID: 2-s2.0-85046135167ISBN: 978-1-5090-2873-3 OAI: oai:DiVA.org:kth-223860DiVA, id: diva2:1187896
Konferanse
IEEE 56th Annual Conference on Decision and Control (CDC), DEC 12-15, 2017, Melbourne, Australia
Forskningsfinansiär
Swedish Research Council, 2016-06079
Merknad

QC 20180306

Tilgjengelig fra: 2018-03-06 Laget: 2018-03-06 Sist oppdatert: 2018-06-01bibliografisk kontrollert

Open Access i DiVA

Fulltekst mangler i DiVA

Andre lenker

Forlagets fulltekstScopus

Personposter BETA

Mattila, RobertRojas, Cristian R.

Søk i DiVA

Av forfatter/redaktør
Mattila, RobertRojas, Cristian R.Wahlberg, Bo
Av organisasjonen

Søk utenfor DiVA

GoogleGoogle Scholar

doi
isbn
urn-nbn

Altmetric

doi
isbn
urn-nbn
Totalt: 7 treff
RefereraExporteraLink to record
Permanent link

Direct link
Referera
Referensformat
  • apa
  • harvard1
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • Annet format
Fler format
Språk
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Annet språk
Fler språk
Utmatningsformat
  • html
  • text
  • asciidoc
  • rtf