Approximate nonparametric maximum likelihood for mixture models: A convex optimization approach to fitting arbitrary multivariate mixing distributions.
Long FengLee H. DickerPublished in: Comput. Stat. Data Anal. (2018)
Keyphrases
- convex optimization
- mixture model
- maximum likelihood
- finite mixtures
- dirichlet process
- em algorithm
- parameter estimation
- density estimation
- expectation maximization
- gaussian mixture model
- exponential family
- likelihood function
- gaussian distribution
- mixture components
- maximum likelihood estimation
- gaussian mixture
- mixture modeling
- probability density function
- total variation
- hyperparameters
- maximum a posteriori
- dirichlet distribution
- approximate inference
- generative model
- computer vision
- variational inference
- denoising
- probabilistic model
- bayesian framework
- cross validation
- unsupervised learning
- model selection
- natural images
- higher order
- reinforcement learning
- multiscale
- minimum message length
- information retrieval