MDL-Based Selection of the Number of Components in Mixture Models for Pattern Classification.
Hiroshi TenmotoMineichi KudoMasaru ShimboPublished in: SSPR/SPR (1998)
Keyphrases
- pattern classification
- mixture model
- mixture components
- em algorithm
- nearest neighbor rule
- density estimation
- number of mixture components
- gaussian mixture model
- pattern recognition
- expectation maximization
- prior knowledge
- generative model
- maximum likelihood
- language model
- probabilistic model
- decision boundary
- object recognition
- finite mixtures
- feature extraction