Theoretical and practical considerations on the convergence properties of the Fisher-EM algorithm.
Charles BouveyronCamille BrunetPublished in: J. Multivar. Anal. (2012)
Keyphrases
- em algorithm
- expectation maximization
- maximum likelihood
- mixture model
- maximum likelihood estimation
- generative model
- likelihood function
- gaussian mixture model
- parameter estimation
- expectation maximisation
- log likelihood
- incomplete data
- hyperparameters
- hidden variables
- gaussian mixture
- mixture distribution
- convergence rate
- probability density function
- parameter learning
- maximum a posteriori
- mixture modeling
- penalized likelihood
- finite mixture model
- image processing
- model selection
- likelihood maximization