Speaker model selection based on the Bayesian information criterion applied to unsupervised speaker indexing.
Masafumi NishidaTatsuya KawaharaPublished in: IEEE Trans. Speech Audio Process. (2005)
Keyphrases
- model selection
- bayesian information criterion
- speaker diarization
- cross validation
- information criterion
- hyperparameters
- unsupervised learning
- selection criterion
- mixture model
- sample size
- parameter estimation
- statistical inference
- marginal likelihood
- machine learning
- model selection criteria
- regression model
- variable selection
- gaussian process
- speech recognition
- feature selection
- learning machines
- parameter determination
- speaker verification
- gaussian mixture model
- posterior distribution
- feature space
- feature extraction
- information retrieval
- data sets