Unsupervised learning via mixtures of skewed distributions with hypercube contours.
Brian C. FranczakCristina TortoraRyan P. BrownePaul D. McNicholasPublished in: Pattern Recognit. Lett. (2015)
Keyphrases
- unsupervised learning
- mixture model
- heavy tailed distributions
- expectation maximization
- highly skewed
- heavy tailed
- supervised learning
- mixture distribution
- exponential family
- active contours
- probability distribution
- dimensionality reduction
- dirichlet distribution
- class distribution
- em algorithm
- master slave
- reinforcement learning
- image processing
- mixtures of gaussians
- parallel algorithm
- power law
- normal distribution
- machine learning
- mixture components
- kl divergence
- gaussian mixture model
- random variables
- semi supervised
- feature selection
- gaussian mixture
- three dimensional
- training data
- factor analyzers
- dirichlet processes
- neural network
- mixture distributions
- face recognition
- contour matching
- pairwise
- model selection
- polygonal approximation
- gaussian distribution
- blind source separation