SampleBoost: Improving boosting performance by destabilizing weak learners based on weighted error analysis.
Mohamed AbouelenienXiaohui YuanPublished in: ICPR (2012)
Keyphrases
- weak learners
- error analysis
- ensemble learning
- multi class classification
- boosting algorithms
- least squares
- adaboost algorithm
- multi class
- weak classifiers
- linear classifiers
- generalization error
- binary classification problems
- error correction
- ensemble methods
- cost sensitive
- data sets
- base classifiers
- binary classifiers
- pairwise
- decision trees
- computer vision