Empirical Study of Individual Feature Evaluators and Cutting Criteria for Feature Selection in Classification.
Antonio Arauzo-AzofraJosé Luis Aznarte-MelladoJosé Manuel BenítezPublished in: ISDA (2009)
Keyphrases
- empirical studies
- feature selection
- feature set
- irrelevant features
- discriminative features
- classification accuracy
- feature vectors
- feature subset
- text classification
- feature ranking
- uci datasets
- evaluation criteria
- redundant features
- feature extraction
- support vector machine
- classification models
- support vector
- empirical analysis
- class separability
- input features
- feature space
- feature weights
- image classification
- pattern recognition
- mutual information
- decision trees
- selection criterion
- feature selection and classification
- high dimensionality
- machine learning
- feature selection algorithms
- feature subset selection
- individual features
- classification performances
- text categorization
- unsupervised learning
- feature reduction
- bayes classifier
- multi class
- supervised feature selection
- real world data sets
- experimental design
- naive bayes
- feature weighting
- accurate classification
- feature values
- single feature
- feature representation
- information gain
- svm classifier
- data pre processing
- class labels
- discriminatory power
- image features
- supervised learning