Totally corrective boosting algorithms that maximize the margin.
Manfred K. WarmuthJun LiaoGunnar RätschPublished in: ICML (2006)
Keyphrases
- boosting algorithms
- multi class
- loss function
- cost sensitive
- soft margin
- binary classification
- adaboost algorithm
- weak learners
- generalization error
- base classifiers
- weak classifiers
- training error
- learning algorithm
- ensemble methods
- bayesian networks
- data sets
- learning problems
- linear combination
- support vector machine
- feature vectors
- high dimensional
- minimum margin