Variability compensation in small data: Oversampled extraction of i-vectors for the classification of depressed speech.
Nicholas CumminsJulien EppsVidhyasaharan SethuJarek KrajewskiPublished in: ICASSP (2014)
Keyphrases
- data sets
- data analysis
- database
- classification accuracy
- data quality
- data collection
- prior knowledge
- feature selection
- data structure
- high quality
- classification models
- original data
- data extraction
- input data
- small number
- image data
- knowledge discovery
- data points
- data sources
- feature extraction
- training data
- decision trees
- data mining techniques
- training samples
- data distribution
- raw data
- xml documents
- data mining
- support vector