Aggregate and mixed-order Markov models for statistical language processing

dc.creatorSaul, Lawrence
dc.creatorPereira, Fernando
dc.date1997-06-09
dc.date.accessioned2026-07-07T09:10:50Z
dc.date.available2026-07-07T09:10:50Z
dc.descriptionWe consider the use of language models whose size and accuracy are intermediate between different order n-gram models. Two types of models are studied in particular. Aggregate Markov models are class-based bigram models in which the mapping from words to classes is probabilistic. Mixed-order Markov models combine bigram models whose predictions are conditioned on different words. Both types of models are trained by Expectation-Maximization (EM) algorithms for maximum likelihood estimation. We examine smoothing procedures in which these models are interposed between different order n-grams. This is found to significantly reduce the perplexity of unseen word combinations.
dc.description9 pages, 4 PostScript figures, uses psfig.sty and aclap.sty; to appear in the proceedings of EMNLP-2
dc.identifierhttps://arxiv.org/abs/cmp-lg/9706007
dc.identifierhttp://arxiv.org/abs/cmp-lg/9706007
dc.identifier.urihttp://salesiana.dossiersoluciones.com/handle/123456789/151471
dc.subjectComputation and Language
dc.titleAggregate and mixed-order Markov models for statistical language processing
dc.typetext

Files

Collections