Multi-scale Sentiment Classification Using Canonical Correlation Analysis on Riemannian Manifolds.
Shuanglu DaiXingzhong XuBitian JiangHong ManPublished in: ISM (2016)
Keyphrases
- sentiment classification
- riemannian manifolds
- multiscale
- euclidean space
- sentiment analysis
- text classification
- cross domain
- multi class classification
- mean shift
- cross lingual
- vector space
- domain adaptation
- feature space
- covariance matrices
- parameter space
- geodesic distance
- image representation
- manifold learning
- geometric structure
- image segmentation
- covariance matrix
- image processing
- feature selection
- shape analysis
- multi class
- high dimensional
- support vector
- reinforcement learning
- bayesian networks
- transfer learning
- labeled data
- low dimensional
- principal component analysis
- feature vectors
- reproducing kernel hilbert space
- data sets