Linear classifier combination and selection using group sparse regularization and hinge loss.
Mehmet Umut SenHakan ErdoganPublished in: Pattern Recognit. Lett. (2013)
Keyphrases
- classifier combination
- hinge loss
- loss function
- risk minimization
- potential functions
- reproducing kernel hilbert space
- convex optimization
- pattern recognition
- binary classification
- probability estimation
- multi class
- pairwise
- base classifiers
- machine learning
- input space
- support vector
- kernel methods
- high dimensional
- neural network
- generalization error
- text classification
- markov random field
- information extraction
- similarity measure