Kernel methods and the exponential family.
Stéphane CanuAlexander J. SmolaPublished in: Neurocomputing (2006)
Keyphrases
- kernel methods
- exponential family
- closed form
- maximum likelihood
- graphical models
- statistical models
- density estimation
- log likelihood
- missing values
- mixture model
- support vector
- feature space
- kernel function
- support vector machine
- order statistics
- machine learning
- variational methods
- hidden variables
- probability density function
- expectation maximization
- parameter estimation
- belief propagation
- markov chain monte carlo
- data sets
- unsupervised learning
- em algorithm