A criterion for model selection using minimum description length.

Amir Najmi Richard A. Olshen Robert M. Gray

Published in: SEQUENCES (1997)

Keyphrases

minimum description length
model selection
cross validation
information theoretic
mdl principle
parameter estimation
hyperparameters
sample size

bayesian learning
statistical learning
selection criterion
machine learning
feature selection
variable selection
model selection criteria
motion segmentation

regression model
meta learning
mixture model
parameter determination
statistical inference
error estimation
generalization error

gaussian process
information criterion
marginal likelihood
generalization bounds
data mining
mutual information
automatic model selection