Probabilistic Spherical Discriminant Analysis: An Alternative to PLDA for length-normalized embeddings.
Niko BrummerAlbert SwartLadislav MosnerAnna SilnovaOldrich PlchotThemos StafylakisLukás BurgetPublished in: INTERSPEECH (2022)
Keyphrases
- discriminant analysis
- linear discriminant analysis
- dimensionality reduction
- face recognition
- principal component analysis
- high dimensional data
- dimension reduction
- feature extraction
- fisher discriminant analysis
- bayesian networks
- kernel discriminant analysis
- generalized linear
- generative model
- low dimensional
- partial least squares
- manifold learning
- graph embedding
- dimensionality reduction methods
- kernel principal component analysis
- vector space
- fisher criterion
- support vector
- fisher linear discriminant analysis
- neural network
- discriminant subspace
- probabilistic model
- similarity measure
- decision trees
- class separability
- linear discriminant
- support vector machine svm
- multi class
- discriminant vectors
- feature selection