Maximum entropy relaxation for multiscale graphical model selection.
Myung Jin ChoiVenkat ChandrasekaranAlan S. WillskyPublished in: ICASSP (2008)
Keyphrases
- maximum entropy
- model selection
- multiscale
- cross validation
- maximum entropy principle
- hyperparameters
- variable selection
- parameter estimation
- regression model
- mixture model
- sample size
- information criterion
- selection criterion
- gaussian process
- image segmentation
- generalization error
- image processing
- conditional random fields
- machine learning
- unsupervised learning
- feature selection
- higher order
- bayesian methods
- feature vectors
- bayesian information criterion
- data mining
- model selection criteria