ACL RD-TEC 1.0 Summarization of W03-0420
Paper Title:
MAXIMUM ENTROPY MODELS FOR NAMED ENTITY RECOGNITION
MAXIMUM ENTROPY MODELS FOR NAMED ENTITY RECOGNITION
Authors: Oliver Bender and Franz Josef Och and Hermann Ney
Primarily assigned technology terms:
- algorithm
- approximation
- capitalization
- count-based feature reduction
- entity recognition
- feature selection
- generalized iterative scaling
- iterative scaling
- maximum approximation
- maximum entropy
- maximum entropy framework
- named entity recognition
- ne recognizer
- optimization
- processing
- recognition
- recognizer
- search
- smoothing
- smoothing method
- training procedure
- viterbi
- viterbi search
Other assigned terms:
- annotated corpus
- annotated training set
- approach
- chunk
- chunk tag
- context information
- convergence
- data sets
- dictionary
- entropy
- entropy models
- feature
- gaussian prior
- generation
- german corpus
- knowledge
- lambda
- language-dependent knowledge
- lexical features
- likelihood
- maximum entropy models
- method
- model parameters
- multilingual text
- named entities
- named entity
- names
- natural language
- nouns
- optimization criterion
- part-of-speech
- pos tag
- posterior
- posterior probability
- prior probability
- probability
- procedure
- process
- proper names
- punctuation
- punctuation marks
- recognition accuracy
- search problem
- search procedure
- sentence
- sentence boundaries
- sentences
- standard deviation
- suffix
- suffixes
- tag sequence
- tags
- test phase
- text
- text corpus
- tokens
- training
- training criterion
- training data
- training set
- vocabulary
- word
- word features
- words
