Direct Importance Estimation with a Mixture of Probabilistic Principal Component Analyzers.
Makoto YamadaMasashi SugiyamaGordon WichernJaak SimmPublished in: IEICE Trans. Inf. Syst. (2010)
Keyphrases
- principal components
- principal component analysis
- dimensionality reduction
- multivariate data
- probabilistic model
- mixture model
- kernel space
- principal component regression
- continuous valued
- generative model
- correlation coefficient
- bayesian networks
- covariance matrix
- generalized em algorithm
- latent variable models
- machine learning
- parameter estimation
- hyperplane
- expectation maximization
- principal curves
- neural network
- posterior probability
- maximum likelihood estimation
- unsupervised learning
- graphical models
- high dimensional
- mixture distributions
- feature extraction