Feature Selection via Probabilistic Outputs.
Andrea Pohoreckyj DanylukNicholas ArnostiPublished in: ICML (2012)
Keyphrases
- feature selection
- text categorization
- data driven
- feature selection algorithms
- feature set
- text classification
- classification models
- information gain
- feature space
- model selection
- multi class
- data sets
- context sensitive
- high dimensionality
- classification accuracy
- probabilistic approaches
- support vector
- bayesian networks
- probability theory
- feature ranking
- small sample
- unsupervised feature selection
- method for feature selection
- feature subset
- microarray data
- information theoretic
- mutual information
- rough sets
- probability distribution
- probabilistic model
- data analysis
- neural network