Cross-validation and aggregated EM training for robust parameter estimation.
Takahiro ShinozakiMari OstendorfPublished in: Comput. Speech Lang. (2008)
Keyphrases
- sample size
- cross validation
- model selection
- robust parameter estimation
- training set
- mixture model
- hyperparameters
- generalization error
- training and test data
- variable selection
- expectation maximization
- grid search
- cross validated
- machine learning
- error estimates
- feature selection
- regularization parameter
- em algorithm
- training examples
- test set
- supervised learning
- density estimation
- active learning
- unseen data
- training data
- prior knowledge
- data sets
- nearest neighbor classifiers
- support vector
- probabilistic model
- nearest neighbor
- training samples