ACL RD-TEC 1.0 Summarization of P06-1028
Paper Title:
TRAINING CONDITIONAL RANDOM FIELDS WITH MULTIVARIATE EVALUATION MEASURES
TRAINING CONDITIONAL RANDOM FIELDS WITH MULTIVARIATE EVALUATION MEASURES
Authors: Jun Suzuki and Erik McDermott and Hideki Isozaki
Primarily assigned technology terms:
- algorithm
- approximation
- binary classification
- chunking
- classification
- classifier
- computational linguistics
- conditional random fields
- crf training
- crfs
- english ner
- entity recognition
- error minimization
- feature selection
- forward-backward algorithm
- l-bfgs optimization
- language learning
- learning
- loss function
- machine translation
- map training
- mce criterion training
- modeling
- named entity recognition
- natural language learning
- normalization
- numerical optimization
- optimization
- optimization procedure
- parameter optimization
- parameter tuning
- recognition
- regularization
- reranking
- segmentation
- smoothing
- statistical machine translation
- support vector machine
- tagging
- text chunking
- text classification
- training procedure
- tuning
- viterbi
- viterbi algorithm
- voting
Other assigned terms:
- accuracy-based evaluation
- approach
- association for computational linguistics
- bayes decision rule
- bigram
- case
- chunk
- classification error
- conditional probability
- data set
- decision rule
- development set
- dictionary
- distribution
- error rate
- evaluation measure
- evaluation measures
- f-score
- fact
- feature
- feature set
- feature vector
- formalism
- formalization
- gaussian prior
- imbalanced data set
- labeling
- likelihood
- linear combination
- linguistics
- log-likelihood
- log-linear combination
- measure
- measures
- method
- named entity
- natural language
- parameter values
- part-of-speech
- part-of-speech tags
- posteriori probability
- prefixes and suffixes
- prior distribution
- probability
- procedure
- process
- regular expressions
- segment-wise discriminant function
- segments
- sentence
- sentences
- sigmoid function
- suffixes
- support vector
- tagging scheme
- tags
- term
- terms
- test data
- text
- tokens
- training
- training and test data
- training criterion
- training data
- window size
- word
- words
