A min-max regret approach to maximum likelihood inference under incomplete data.
Romain GuillaumeDidier DuboisPublished in: Int. J. Approx. Reason. (2020)
Keyphrases
- min max
- incomplete data
- maximum likelihood
- em algorithm
- bayesian networks
- hyperparameters
- learning bayesian networks
- expectation maximization
- parameter learning
- missing data
- max min
- parameter estimation
- missing values
- online learning
- maximum a posteriori
- lower bound
- maximum likelihood estimation
- generative model
- incomplete data sets
- bayes classifier
- exponential family
- multiple imputation
- variational bayesian
- typical testors
- bayesian inference
- markov networks
- belief networks
- hidden variables
- image processing
- markov chain monte carlo
- bayesian model
- machine learning
- input data
- principal component analysis
- variational inference
- nearest neighbor
- irrelevant attributes
- image segmentation
- computer vision