Weakly Supervised Feature Compression Based Topic Model for Sentiment Classification.
Yan HuXiaofei XuLi LiPublished in: KSEM (2017)
Keyphrases
- weakly supervised
- topic models
- sentiment classification
- text classification
- topic modeling
- latent dirichlet allocation
- text mining
- text documents
- co occurrence
- probabilistic model
- image features
- generative model
- cross lingual
- object class
- superpixels
- text categorization
- knowledge base
- training data
- bayesian networks
- feature selection
- feature set
- object detection
- knowledge discovery