A training method for average voice model based on shared decision tree context clustering and speaker adaptive training.
Junichi YamagishiTakashi MasukoKeiichi TokudaTakao KobayashiPublished in: ICASSP (1) (2003)
Keyphrases
- clustering method
- training process
- training phase
- training algorithm
- training set
- feed forward neural networks
- significant improvement
- unsupervised learning
- training samples
- decision trees
- cluster analysis
- detection method
- learning algorithm
- similarity function
- spectral clustering
- supervised learning
- classification accuracy
- pairwise
- preprocessing
- mutual information
- logistic regression
- high accuracy
- probabilistic model
- data sets
- decision tree classifiers