Matches in SemOpenAlex for { <https://semopenalex.org/work/W1581013037> ?p ?o ?g. }
- W1581013037 abstract "Viterbi Training (VT) provides a fast but inconsistent estimator of Hidden Markov Models (HMM). The inconsistency is alleviated with little extra computation when we enable VT to asymptotically fix the true values of the parameters. This relies on infinite Viterbi alignments and associated with them limiting probability distributions. First in a sequel, this paper is a proof of concept; it focuses on mixture models, an important but special case of HMM where the limiting distributions can be calculated exactly. A simulated Gaussian mixture shows that our central algorithm (VA1) can significantly improve the accuracy of VT with little extra cost. Next in the sequel, we present elsewhere a theory of the adjusted VT for the general HMMs, where the limiting distributions are more challenging to find. Here, we also present another, more advanced correction to VT, verify its fast convergence and high accuracy; its computational feasibility requires additional investigation." @default.
- W1581013037 created "2016-06-24" @default.
- W1581013037 creator A5060111294 @default.
- W1581013037 creator A5079016144 @default.
- W1581013037 date "2005-01-01" @default.
- W1581013037 modified "2023-10-03" @default.
- W1581013037 title "Adjusted Viterbi Training : a proof of concept" @default.
- W1581013037 cites W1486632395 @default.
- W1581013037 cites W1498225184 @default.
- W1581013037 cites W1508165687 @default.
- W1581013037 cites W1511268987 @default.
- W1581013037 cites W1560013842 @default.
- W1581013037 cites W1579271636 @default.
- W1581013037 cites W1968149043 @default.
- W1581013037 cites W1973923101 @default.
- W1581013037 cites W1985093013 @default.
- W1581013037 cites W1988506022 @default.
- W1581013037 cites W2007321142 @default.
- W1581013037 cites W2011832962 @default.
- W1581013037 cites W2028486550 @default.
- W1581013037 cites W2029252813 @default.
- W1581013037 cites W2059279601 @default.
- W1581013037 cites W2071516626 @default.
- W1581013037 cites W2074136837 @default.
- W1581013037 cites W2086943813 @default.
- W1581013037 cites W2095589166 @default.
- W1581013037 cites W2106994558 @default.
- W1581013037 cites W2110155683 @default.
- W1581013037 cites W2111724331 @default.
- W1581013037 cites W2120283843 @default.
- W1581013037 cites W2125838338 @default.
- W1581013037 cites W2131119226 @default.
- W1581013037 cites W2131578006 @default.
- W1581013037 cites W2139516179 @default.
- W1581013037 cites W2142775654 @default.
- W1581013037 cites W2151443911 @default.
- W1581013037 cites W2161562001 @default.
- W1581013037 cites W2166866346 @default.
- W1581013037 cites W2615953416 @default.
- W1581013037 cites W2896749479 @default.
- W1581013037 cites W3101583054 @default.
- W1581013037 cites W3148186152 @default.
- W1581013037 cites W2072634211 @default.
- W1581013037 hasPublicationYear "2005" @default.
- W1581013037 type Work @default.
- W1581013037 sameAs 1581013037 @default.
- W1581013037 citedByCount "8" @default.
- W1581013037 countsByYear W15810130372018 @default.
- W1581013037 countsByYear W15810130372021 @default.
- W1581013037 crossrefType "journal-article" @default.
- W1581013037 hasAuthorship W1581013037A5060111294 @default.
- W1581013037 hasAuthorship W1581013037A5079016144 @default.
- W1581013037 hasConcept C105795698 @default.
- W1581013037 hasConcept C113638808 @default.
- W1581013037 hasConcept C11413529 @default.
- W1581013037 hasConcept C119857082 @default.
- W1581013037 hasConcept C121332964 @default.
- W1581013037 hasConcept C127413603 @default.
- W1581013037 hasConcept C130319729 @default.
- W1581013037 hasConcept C154945302 @default.
- W1581013037 hasConcept C157125643 @default.
- W1581013037 hasConcept C162324750 @default.
- W1581013037 hasConcept C163716315 @default.
- W1581013037 hasConcept C163836022 @default.
- W1581013037 hasConcept C185429906 @default.
- W1581013037 hasConcept C188198153 @default.
- W1581013037 hasConcept C193969084 @default.
- W1581013037 hasConcept C196455857 @default.
- W1581013037 hasConcept C23224414 @default.
- W1581013037 hasConcept C2777303404 @default.
- W1581013037 hasConcept C28490314 @default.
- W1581013037 hasConcept C33923547 @default.
- W1581013037 hasConcept C41008148 @default.
- W1581013037 hasConcept C45374587 @default.
- W1581013037 hasConcept C50522688 @default.
- W1581013037 hasConcept C54907487 @default.
- W1581013037 hasConcept C57273362 @default.
- W1581013037 hasConcept C60582962 @default.
- W1581013037 hasConcept C61224824 @default.
- W1581013037 hasConcept C62520636 @default.
- W1581013037 hasConcept C78519656 @default.
- W1581013037 hasConcept C98763669 @default.
- W1581013037 hasConceptScore W1581013037C105795698 @default.
- W1581013037 hasConceptScore W1581013037C113638808 @default.
- W1581013037 hasConceptScore W1581013037C11413529 @default.
- W1581013037 hasConceptScore W1581013037C119857082 @default.
- W1581013037 hasConceptScore W1581013037C121332964 @default.
- W1581013037 hasConceptScore W1581013037C127413603 @default.
- W1581013037 hasConceptScore W1581013037C130319729 @default.
- W1581013037 hasConceptScore W1581013037C154945302 @default.
- W1581013037 hasConceptScore W1581013037C157125643 @default.
- W1581013037 hasConceptScore W1581013037C162324750 @default.
- W1581013037 hasConceptScore W1581013037C163716315 @default.
- W1581013037 hasConceptScore W1581013037C163836022 @default.
- W1581013037 hasConceptScore W1581013037C185429906 @default.
- W1581013037 hasConceptScore W1581013037C188198153 @default.
- W1581013037 hasConceptScore W1581013037C193969084 @default.
- W1581013037 hasConceptScore W1581013037C196455857 @default.
- W1581013037 hasConceptScore W1581013037C23224414 @default.
- W1581013037 hasConceptScore W1581013037C2777303404 @default.