Minimum description length vs. maximum likelihood in lossy data compression.
Mokshay MadimanMatthew T. HarrisonIoannis KontoyiannisPublished in: ISIT (2004)
Keyphrases
- minimum description length
- maximum likelihood
- information theoretic
- mdl principle
- parameter estimation
- expectation maximization
- likelihood function
- em algorithm
- scoring metrics
- maximum likelihood estimator
- gaussian mixture model
- log likelihood
- maximum likelihood estimation
- maximum a posteriori
- parameter learning
- additive noise
- gaussian distribution
- information theory
- data mining
- belief propagation
- generative model
- least squares
- image segmentation