Gaussian mixture optimization for HMM based on efficient cross-validation.
Takahiro ShinozakiTatsuya KawaharaPublished in: INTERSPEECH (2007)
Keyphrases
- cross validation
- gaussian mixture
- model selection
- hyperparameters
- support vector
- generalization error
- grid search
- cross validated
- mixtures of gaussians
- density estimation
- mixture model
- probability density function
- worst case
- hidden markov models
- training set
- decision trees
- sample size
- covariance matrix
- nearest neighbor
- upper bound
- ls svm
- unseen data