ACL RD-TEC 1.0 Summarization of W02-0811
Paper Title:
COMBINING HETEROGENEOUS CLASSIFIERS FOR WORD SENSE DISAMBIGUATION
COMBINING HETEROGENEOUS CLASSIFIERS FOR WORD SENSE DISAMBIGUATION
Authors: Dan Klein and Kristina Toutanova and H. Tolga Ilhan and Sepandar D. Kamvar and Christopher D. Manning
Primarily assigned technology terms:
- algorithm
- bootstrap
- classification
- classification algorithm
- classifier
- classifier combination
- classifier construction
- classifiers
- computational linguistics
- cross-validation
- decision trees
- disambiguation
- entropy classifier
- global selection
- language processing
- majority voting
- maximum entropy
- maximum entropy approach
- maximum entropy classifier
- maximum entropy model
- natural language processing
- processing
- ranking
- reporting
- selection method
- sense disambiguation
- simple majority voting
- single classifier
- smoothing
- statistical estimation
- supervised word sense disambiguation
- training procedure
- voting
- weighted voting
- weighting
- word sense disambiguation
- word-sense disambiguation
Other assigned terms:
- ambiguous word
- approach
- association for computational linguistics
- case
- citation
- context window
- context words
- convergence
- corpora
- correlation
- data set
- entropy
- estimation
- exponential model
- fact
- feature
- implementation
- likelihood
- linguistics
- measure
- measures
- method
- most-frequent-sense baseline
- n-gram
- n-gram models
- natural language
- oracle
- part-of-speech
- posterior
- posterior probability
- probability
- procedure
- process
- rank correlation
- spearman rank correlation
- stem
- stress
- target word
- test data
- test set
- tokens
- training
- training corpora
- training data
- trees
- vector space
- weighting scheme
- window size
- word
- word sense
- word type
- words
