Conditional leaving-one-out and cross-validation for discount estimation in Kneser-Ney-like extensions.
Jesús Andrés-FerrerMartin SundermeyerHermann NeyPublished in: ICASSP (2012)
Keyphrases
- cross validation
- language model
- model selection
- error estimates
- support vector
- hyperparameters
- information retrieval
- probabilistic model
- variable selection
- parameter estimation
- cross validated
- generalization error
- training set
- regression problems
- mixture model
- regularization parameter
- error estimation
- density estimation
- ls svm
- database
- binary classifiers
- unseen data
- neural network