Classifying With AdaBoost.M1: The Training Error Threshold Myth.
Antônio LeãesPaulo FernandesLucelene LopesJoaquim AssunçãoPublished in: FLAIRS Conference (2017)
Keyphrases
- training error
- adaboost algorithm
- error rate
- generalization error
- boosting algorithms
- classification error
- hidden layer
- face detection
- base classifiers
- multiclass classification
- gradient method
- training data
- learning algorithm
- minimum margin
- cost sensitive
- loss function
- model selection
- active learning
- decision trees
- prediction error
- recurrent neural networks
- error bounds
- training process
- object detection
- upper bound
- machine learning