Asymptotically minimax regret by Bayes mixtures for non-exponential families.
Jun'ichi TakeuchiAndrew R. BarronPublished in: ITW (2013)
Keyphrases
- exponential family
- minimax regret
- mixture model
- preference elicitation
- utility function
- density estimation
- decision problems
- graphical models
- maximum likelihood
- statistical models
- log likelihood
- closed form
- probabilistic model
- em algorithm
- stochastic programming
- decision trees
- gaussian mixture model
- generative model
- missing values
- model selection
- sample size
- hidden variables
- language model
- probability density function
- misclassification costs
- expectation maximization
- support vector machine
- variational methods
- support vector
- order statistics
- unsupervised learning
- learning algorithm
- training data
- class distribution
- random variables
- data sets