Dependency Tree-based Sentiment Classification using CRFs with Hidden Variables.
Tetsuji NakagawaKentaro InuiSadao KurohashiPublished in: HLT-NAACL (2010)
Keyphrases
- hidden variables
- sentiment classification
- conditional random fields
- probabilistic model
- generative model
- markov networks
- sentiment analysis
- cross domain
- text classification
- bayesian networks
- bayesian inference
- latent variables
- opinion mining
- higher order
- graphical models
- em algorithm
- cross lingual
- information extraction
- hidden markov models
- pairwise
- named entity recognition
- missing values
- incomplete data
- knowledge transfer
- bayesian framework
- data mining
- information retrieval
- sentiment polarity
- text categorization
- expectation maximization
- language model
- knowledge discovery
- feature selection
- learning algorithm