Constructing Boosting Algorithms from SVMs: An Application to One-Class Classification.
Gunnar RätschSebastian MikaBernhard SchölkopfKlaus-Robert MüllerPublished in: IEEE Trans. Pattern Anal. Mach. Intell. (2002)
Keyphrases
- boosting algorithms
- binary classification
- support vector
- soft margin
- multi class
- cost sensitive
- early stopping
- multi class classification
- generalization error
- weak learners
- support vector machine
- loss function
- learning problems
- binary classifiers
- multi label
- prediction accuracy
- base classifiers
- binary classification problems
- class imbalance
- ensemble methods
- cross validation
- misclassification costs
- feature selection
- linear classifiers
- generalization ability
- class distribution
- text categorization
- model selection
- object detection
- knn
- upper bound
- learning tasks
- support vector machine svm
- active learning
- pairwise