An alternative to standard maximum likelihood for Gaussian mixtures.
Frédéric ChampagnatJérôme IdierPublished in: ICASSP (1995)
Keyphrases
- gaussian mixture
- maximum likelihood
- expectation maximization
- em algorithm
- covariance matrices
- gaussian mixture model
- mixture model
- parameter estimation
- closed form
- covariance matrix
- probability density function
- maximum likelihood estimation
- high dimensional
- unsupervised learning
- gaussian distribution
- kl divergence