Feature Selection Based on a New Formulation of the Minimal-Redundancy-Maximal-Relevance Criterion.
Daniel PonsaAntonio M. LópezPublished in: IbPRIA (1) (2007)
Keyphrases
- feature selection
- minimum redundancy
- feature relevance
- text categorization
- feature selection algorithms
- support vector
- supervised feature selection
- class separability
- text classification
- minimal length
- mutual information
- selection criterion
- unsupervised feature selection
- high dimensionality
- feature set
- feature space
- relevance feedback
- knn
- selected features
- irrelevant features
- machine learning
- method for feature selection
- multi task
- information retrieval
- discriminative features
- selecting relevant features
- integer points
- data sets
- optimization criterion
- information gain
- microarray data
- unsupervised learning
- multi class
- kernel learning
- integer program
- retrieved documents
- feature subset
- test collection
- dimensionality reduction
- support vector machine
- probabilistic model
- feature extraction