A maximum entropy approach for integrating semantic information in statistical language models.
Chuang-Hua ChuehJen-Tzung ChienHsin-Min WangPublished in: ISCSLP (2004)
Keyphrases
- maximum entropy
- semantic information
- statistical language models
- maximum entropy principle
- wordnet
- language model
- keywords
- domain knowledge
- high level
- low level features
- metadata
- semantic similarity
- semantic features
- retrieval model
- xml documents
- named entity recognition
- conditional random fields
- low level
- vector space model
- contextual information
- higher order
- information retrieval
- language modeling
- query topic