Recover the spectrum of covariance matrix: a non-asymptotic iterative method.
Juntao DuanIonel PopescuHeinrich MatzingerPublished in: CoRR (2022)
Keyphrases
- covariance matrix
- covariance matrices
- principal component analysis
- sample size
- mahalanobis distance
- geometrical interpretation
- gaussian mixture
- positive definite
- eigendecomposition
- eigenvalues and eigenvectors
- multivariate gaussian
- transformation matrix
- estimation error
- symmetric matrix
- objective function
- class conditional densities
- machine learning
- multivariate normal
- correlation matrix
- principal components
- singular value decomposition
- worst case
- probabilistic model
- face recognition
- feature selection