A Convex Surrogate Operator for General Non-Modular Loss Functions.
Jiaqian YuMatthew B. BlaschkoPublished in: CoRR (2016)
Keyphrases
- loss function
- convex loss functions
- risk minimization
- bregman divergences
- empirical risk
- hinge loss
- support vector
- learning to rank
- pairwise
- squared error
- special case
- logistic regression
- boosting algorithms
- reproducing kernel hilbert space
- cost sensitive
- information theoretic
- update rules
- loss minimization
- precision and recall
- ranking functions
- line search