Heteroscedastic Gaussian based correction term for fisher discriminant analysis and its kernel extension.
Tatsuya YokotaToru WakaharaYukihiko YamashitaPublished in: IJCNN (2013)
Keyphrases
- fisher discriminant analysis
- kernel trick
- multiple kernel learning
- discriminant analysis
- linear discriminant analysis
- kernel methods
- gaussian distribution
- gaussian kernel
- maximum margin
- kernel function
- dimensionality reduction
- factor analysis
- dimensionality reduction methods
- principal component analysis
- kernel learning
- input space
- kernel pca
- feature space
- regression algorithm
- linear combination
- computer vision
- support vector
- generative model
- pattern recognition
- kernel matrix
- discriminative learning
- preprocessing step
- locality preserving projections
- machine learning
- sparse coding
- locality preserving