ACL RD-TEC 1.0 Summarization of W04-3235
Paper Title:
ERROR MEASURES AND BAYES DECISION RULES REVISITED WITH APPLICATIONS TO POS TAGGING
ERROR MEASURES AND BAYES DECISION RULES REVISITED WITH APPLICATIONS TO POS TAGGING
Authors: Hermann Ney and Maja Popović and David Sündermann
Primarily assigned technology terms:
- algorithm
- bootstrap
- classification
- coding
- dynamic programming
- dynamic programming algorithm
- entropy modelling
- forward-backward algorithm
- hidden markov
- hidden markov model
- isolated word recognition
- language modelling
- language processing
- loss function
- machine translation
- markov model
- maximum entropy
- maximum entropy method
- maximum entropy model
- maximum likelihood
- modelling
- natural language processing
- nlp
- pattern classification
- pos tagging
- processing
- programming algorithm
- recognition
- smoothing
- speech recognition
- statistical approaches
- statistical methods
- statistical nlp
- tagger
- tagging
- terminology
- training procedure
- viterbi
- viterbi algorithm
- word recognition
Other assigned terms:
- adjective
- adverb
- approach
- bayes decision rule
- bleu
- case
- concepts
- conditional probabilities
- decision rule
- decision theory
- derivation
- determiner
- determiners
- distribution
- edit distance
- entropy
- entropy models
- error rate
- experimental results
- feature
- generative model
- german corpus
- joint distribution
- knowledge
- likelihood
- maximum entropy models
- measure
- measures
- method
- mutual information
- n-gram
- natural language
- nlp task
- nlp tasks
- partof-speech
- past participle
- pos tag
- pos tag sequence
- posterior
- posterior probability
- preposition
- probabilities
- probability
- probability distribution
- probability model
- procedure
- punctuation
- punctuation marks
- sentence
- sentences
- statistical approach
- statistical decision theory
- statistics
- symbol
- symbols
- tag sequence
- tagging task
- tags
- test data
- text
- theory
- tokens
- training
- training corpus
- training criterion
- training data
- trigram
- trigram model
- verb
- vocabulary
- vocabulary size
- word
- word error rate
- words
- wsj corpus