Information-theoretic approaches to SVM feature selection for metagenome read classification.
Elaine GarbarineJoseph DePasqualeVinay GadiaRobi PolikarGail L. RosenPublished in: Comput. Biol. Chem. (2011)
Keyphrases
- information theoretic
- feature selection
- mutual information
- support vector
- support vector machine
- information theory
- support vector machine svm
- classification accuracy
- text classification
- jensen shannon divergence
- feature selection algorithms
- distributional clustering
- theoretic framework
- feature space
- feature set
- text categorization
- classification performances
- svm classifier
- classification method
- feature extraction
- multi modality
- entropy measure
- information bottleneck
- machine learning
- kullback leibler divergence
- multi class
- information theoretic measures
- feature vectors
- machine learning algorithms
- model selection
- multiple kernel learning
- text classifiers
- fold cross validation
- log likelihood
- supervised learning
- jensen shannon
- image segmentation
- training data
- relative entropy
- training set
- image analysis
- feature subset
- classification algorithm
- kernel function