Efficient estimation of Class A noise parameters via the EM algorithm.
Serena M. ZabinH. Vincent PoorPublished in: IEEE Trans. Inf. Theory (1991)
Keyphrases
- em algorithm
- maximum likelihood estimation
- expectation maximization
- parameter estimation
- maximum likelihood
- log likelihood function
- likelihood function
- mixture model
- gaussian mixture
- update equations
- hyperparameters
- maximum likelihood estimates
- expectation maximisation
- gaussian mixture model
- parameter learning
- map estimation
- generative model
- probability density function
- density estimation
- model selection
- probabilistic model
- maximum a posteriori
- mixture of gaussians
- mixture distribution
- image segmentation
- log likelihood
- model based clustering
- likelihood maximization
- mixture modeling
- hidden variables
- incomplete data
- noise level
- class labels
- unsupervised learning
- linear models
- bayesian framework
- closed form
- finite mixture model
- multi modal
- hidden markov models
- pairwise
- feature selection