An empirical comparison of EM, SEM and MCMC performance for problematic Gaussian mixture likelihoods.
José G. DiasMichel WedelPublished in: Stat. Comput. (2004)
Keyphrases
- gaussian mixture
- expectation maximization
- em algorithm
- posterior probability
- likelihood function
- maximum likelihood
- probability density function
- generative model
- markov chain monte carlo
- posterior distribution
- mixture model
- gaussian mixture model
- parameter estimation
- mixtures of gaussians
- mixture of gaussians
- density estimation
- covariance matrix
- closed form
- gaussian densities
- probability density
- probabilistic model
- maximum likelihood estimation
- bayesian framework
- gaussian density
- maximum a posteriori
- gaussian distribution
- random variables
- objective function
- unsupervised learning
- language model
- higher order
- active learning