Unknown-multiple signal source clustering problem using ergodic HMM and applied to speaker classification.
Jin'ichi MurakamiMasahide SugiyamaHideyuki WatanabePublished in: ICSLP (1996)
Keyphrases
- hidden markov models
- unsupervised learning
- classification accuracy
- support vector
- speech recognition
- classification method
- supervised learning
- pattern recognition
- k means
- clustering analysis
- unsupervised clustering
- unsupervised classification
- clustering method
- pattern classification
- supervised classification
- classification algorithm
- self organizing maps
- feature selection
- feature vectors
- feature extraction
- class labels
- frequency domain
- support vector machine svm
- image classification
- signal processing
- high dimensionality
- multiple sources
- feature space
- decision trees
- machine learning