The EM algorithm with gradient function update for discrete mixtures with known (fixed) number of components.
Dankmar BöhningPublished in: Stat. Comput. (2003)
Keyphrases
- em algorithm
- fixed number
- expectation maximization
- mixture model
- mixture components
- gaussian mixture
- finite number
- log likelihood function
- maximum likelihood
- gaussian mixture model
- mixture distribution
- parameter estimation
- generative model
- probability density function
- log likelihood
- unsupervised learning
- likelihood function
- maximum likelihood estimation
- expectation maximisation
- density estimation
- maximum a posteriori
- likelihood maximization
- model based clustering
- penalized likelihood
- mixture modeling
- finite mixtures
- incomplete data
- mixture of gaussians
- finite mixture model
- hidden variables
- probabilistic model
- discrete data
- k means
- maximum likelihood estimates
- gradient method
- density function
- finite mixture models
- least squares