ACL RD-TEC 1.0 Summarization of W03-0407
Paper Title:
BOOTSTRAPPING POS-TAGGERS USING UNLABELLED DATA
BOOTSTRAPPING POS-TAGGERS USING UNLABELLED DATA
Authors: Stephen Clark and James Curran and Miles Osborne
Primarily assigned technology terms:
- agreement-based co-training
- agreement-based co-training method
- agreement-based selection
- algorithm
- bootstrapping
- bootstrapping process
- classification
- classifier
- classifiers
- co-training
- entity classification
- error reduction
- example selection
- exhaustive search
- generalised iterative scaling
- greedy algorithm
- iterative scaling
- learning
- markov model
- maximum entropy
- named entity classification
- parsers
- re-training
- search
- searching
- selection method
- selection process
- self-training
- statistical parsers
- supervised bootstrapping
- tagger
- taggers
- tagging
Other assigned terms:
- agreement rate
- american news corpus
- approach
- cache
- case
- classification task
- conditional model
- conditional probabilities
- entropy
- fact
- generalisation
- implementation
- independence assumption
- labeled training data
- measure
- method
- named entity
- news corpus
- noise
- parameter settings
- part-of-speech
- penn treebank
- probabilities
- process
- seed
- sentence
- sentences
- set size
- suffix
- tag sequence
- tagging performance
- tagging problem
- tags
- target word
- technique
- text
- training
- training data
- training material
- training set
- transition probabilities
- treebank
- word
- words
