Perturbation-Based Eigenvector Updates for On-Line Principal Components Analysis and Canonical Correlation Analysis.
Anant HegdeJosé Carlos PríncipeDeniz ErdogmusUmut OzertemYadunandana N. RaoHemanth PeddaneniPublished in: J. VLSI Signal Process. (2006)
Keyphrases
- principal components analysis
- canonical correlation analysis
- dimensionality reduction methods
- covariance matrix
- multivariate statistical
- dimensionality reduction
- principal component analysis
- canonical correlations
- least squares
- linear discriminant analysis
- principal components
- feature extraction
- partial least squares
- spectral clustering
- sample size
- multivariate statistical analysis
- data sets
- linear combination
- equivalence relationship
- neural network
- singular value decomposition
- support vector machine
- multi class
- lower bound
- feature space
- objective function
- face recognition
- image processing
- computer vision
- kernel cca