Choosing starting values for the EM algorithm for getting the highest likelihood in multivariate Gaussian mixture models.
Christophe BiernackiGilles CeleuxGérard GovaertPublished in: Comput. Stat. Data Anal. (2003)
Keyphrases
- gaussian mixture model
- em algorithm
- maximum likelihood
- expectation maximization
- mixture model
- likelihood function
- finite mixtures
- finite mixture models
- log likelihood
- parameter estimation
- log likelihood function
- mixture of gaussians
- maximum likelihood estimation
- generative model
- maximum likelihood estimates
- gaussian mixture
- probability density function
- gaussian distribution
- unsupervised learning
- hyperparameters
- incomplete data
- multivariate data
- density estimation
- mixture modeling
- feature vectors
- model based clustering
- mixture components
- bayesian framework
- maximum a posteriori
- probabilistic model