Regularization parameter estimation for spectral regression discriminant analysis based on perturbation theory.
Jie GuiZhenan SunTieniu TanPublished in: ICPR (2012)
Keyphrases
- discriminant analysis
- parameter estimation
- perturbation theory
- model selection
- partial least squares
- estimation problems
- generalized linear
- maximum likelihood
- reproducing kernel hilbert space
- regression model
- face recognition
- least squares
- statistical models
- linear discriminant analysis
- maximal cliques
- linear regression
- markov random field
- em algorithm
- random fields
- feature extraction
- posterior distribution
- principal component analysis
- parameter estimation algorithm
- approximate inference
- regression algorithm
- canonical correlation analysis
- feature space
- hyperparameters
- cross validation
- multispectral images
- structure learning
- data mining
- generative model
- neural network