Minimum description length principle for maximum entropy model selection.

Gaurav Pandey Ambedkar Dukkipati

Published in: ISIT (2013)

Keyphrases

maximum entropy
model selection
minimum description length principle
cross validation
maximum entropy principle
hyperparameters
regression model

parameter estimation
machine learning
sample size
mixture model
conditional random fields
variable selection
model selection criteria

generalization error
selection criterion
bayesian information criterion
marginal likelihood
feature space
information criterion
bayesian methods

level set
likelihood function
gaussian process
training data
feature selection
unsupervised learning