2N-th Order Ergodic Multigram HMM for Modeling of Languages without Marked Word Boundariesconsiders segmentation ambiguities and integrates this with a Ngram model, and able to be trained and tested on a raw, unsegmented...
3Disambiguation of Super Parts of Speech (or Supertags): Allnost Parsingthat is sensitive to context is the n-gram model. The ngram model takes into account the eontextuaJ dependency probabilities...
4Bayesian Semi-Supervised Chinese Word Segmentation for Statistical Machine Translation(Press et al., 2002). The language model is a statistical ngram model estimated using modified Kneser-Ney smoothing. 3 Unigram...