ACL RD-TEC 1.0 Summarization of W99-0617

Paper Title:
POS TAGS AND DECISION TREES FOR LANGUAGE MODELING

Other assigned terms:

  • acoustic model
  • acoustic models
  • acoustic signal
  • adjective
  • approach
  • backoff
  • backoff model
  • bigram
  • bigram model
  • binary tree
  • class-based approach
  • class-based model
  • classification tree
  • cluster
  • conditional probability
  • contextual information
  • decision tree model
  • derivation
  • distribution
  • entropy
  • error rate
  • estimation
  • fact
  • interpolation
  • interpretation
  • knowledge
  • language model
  • language model probability
  • language models
  • large vocabulary speech
  • leaf
  • lexical information
  • likelihood
  • linguistic
  • linguistic knowledge
  • lob corpus
  • maps
  • measure
  • memory space
  • model probability
  • model size
  • mutual information
  • n-gram
  • n-gram language model
  • natural language
  • nouns
  • penn treebank
  • perplexity
  • perplexity measure
  • perplexity reduction
  • personal pronouns
  • pos information
  • pos sequence
  • pos tag
  • pos-based model
  • probabilities
  • probability
  • probability distribution
  • probability distributions
  • probability estimates
  • procedure
  • process
  • pronouns
  • relative frequency
  • root node
  • sentence
  • signal
  • sources of information
  • speech recognition performance
  • speech recognition problem
  • syntactic information
  • syntactic knowledge
  • tags
  • tagset
  • technique
  • term
  • terms
  • test corpus
  • test data
  • test set
  • tokens
  • toolkit
  • training
  • training corpus
  • training data
  • transcripts
  • tree
  • tree model
  • treebank
  • trees
  • trigram
  • trigram model
  • unigram
  • verb
  • vocabulary
  • vocabulary size
  • wall street journal corpus
  • wilcoxon test
  • word
  • word classes
  • word error rate
  • word fragments
  • word sequence
  • word-based model
  • words
  • z-score

Extracted Section Types:


This page last edited on 10 May 2017.

*** ***