HMM Revises Low Marginal Probability by CRF for Chinese Word Segmentation.
Degen HuangDeqin TongYanyan LuoPublished in: CIPS-SIGHAN (2010)
Keyphrases
- chinese word segmentation
- conditional random fields
- hidden markov models
- loopy belief propagation
- marginal probabilities
- graphical models
- pos tagging
- word segmentation
- upper and lower bounds
- pairwise
- maximum entropy
- random fields
- named entity recognition
- natural language understanding
- belief propagation
- joint distribution
- semi supervised learning
- markov random field
- higher order
- information extraction
- probabilistic model
- learning algorithm