On the estimation of a covariance matrix in designing Parzen classifiers.
Yoshihiko HamamotoYasushi FujimotoShingo TomitaPublished in: Pattern Recognit. (1996)
Keyphrases
- covariance matrix
- covariance matrices
- estimation error
- class conditional densities
- density estimation
- principal component analysis
- sample size
- support vector
- mahalanobis distance
- training data
- eigendecomposition
- positive definite
- linear classifiers
- gaussian mixture
- training set
- multivariate gaussian
- geometrical interpretation
- non stationary
- cma es
- feature selection
- principal components
- objective function
- svm classifier
- pseudo inverse
- eigenvalues and eigenvectors
- symmetric matrix
- feature set
- high dimensional
- mahalanobis metric