ACL RD-TEC 1.0 Summarization of P04-3024
Paper Title:
A NEW FEATURE SELECTION SCORE FOR MULTINOMIAL NAIVE BAYES TEXT CLASSIFICATION BASED ON KL-DIVERGENCE
A NEW FEATURE SELECTION SCORE FOR MULTINOMIAL NAIVE BAYES TEXT CLASSIFICATION BASED ON KL-DIVERGENCE
Primarily assigned technology terms:
- 5-fold cross-validation
- approximation
- bayes classifier
- bayes text classification
- binary classifier
- classification
- classifier
- computing
- cross-validation
- document generation
- feature selection
- language processing
- learning
- learning technique
- likelihood estimation
- machine learning
- maximum likelihood
- maximum likelihood estimation
- naive bayes
- naive bayes classifier
- natural language processing
- processing
- scoring
- scoring function
- scoring method
- single classifier
- text classification
Other assigned terms:
- analogy
- british national corpus
- case
- characters
- classification accuracy
- classification error
- data sets
- distribution
- document
- entropy
- estimation
- feature
- generation
- kullback-leibler divergence
- language processing tasks
- likelihood
- measure
- measures
- method
- model parameters
- multinomial model
- mutual information
- natural language
- natural language processing tasks
- precision
- probabilities
- probability
- probability distribution
- processing tasks
- statistics
- stochastic model
- technique
- terms
- text
- text documents
- topics
- training
- training documents
- user
- vocabulary
- vocabulary size
- web content
- word
- word distribution
- words
