On the intrinsic robustness of some leading classifiers and symetric loss function - an empirical evaluation.
Hugo Le BaherVincent LemaireRomain TrinquartPublished in: CoRR (2020)
Keyphrases
- loss function
- support vector
- boosting framework
- boosting algorithms
- pairwise
- logistic regression
- learning to rank
- empirical risk
- risk minimization
- svm classifier
- training data
- reproducing kernel hilbert space
- decision trees
- feature selection
- training set
- stochastic gradient descent
- support vector machine
- decision boundary
- solution path
- regularization term
- linear classifiers
- naive bayes
- cross validation
- generalization ability
- class labels
- feature set
- machine learning methods
- pairwise constraints
- classification algorithm
- active learning
- gaussian kernels
- kernel function
- machine learning algorithms