An improvement of the NEC criterion for assessing the number of clusters in a mixture model.
Christophe BiernackiGilles CeleuxGérard GovaertPublished in: Pattern Recognit. Lett. (1999)
Keyphrases
- mixture model
- gaussian mixture model
- bayesian information criterion
- clustering algorithm
- em algorithm
- model selection
- generative model
- unsupervised learning
- mixture modeling
- probabilistic model
- language model
- mixture components
- probability density function
- highly correlated
- number of mixture components
- data mining
- maximum likelihood
- hidden markov models
- learning algorithm
- model based clustering