Using Random Forests for Prosodic Break Prediction Based on Automatic Speech Labeling.
Olga KhomitsevichPavel ChistikovDmitriy ZakharovPublished in: SPECOM (2014)
Keyphrases
- random forests
- speech recognition
- ensemble methods
- prediction accuracy
- random forest
- logistic regression
- text to speech synthesis
- decision trees
- text to speech
- prosodic features
- machine learning algorithms
- decision tree ensembles
- tree ensembles
- speech synthesis
- active learning
- randomized trees
- regression forests
- conformal prediction
- hidden markov models
- speech signal
- neural network
- head pose estimation
- feature space
- multi class
- genetic programming