Risk Bounds for the Majority Vote: From a PAC-Bayesian Analysis to a Learning Algorithm.
Pascal GermainAlexandre LacasseFrançois LavioletteMario MarchandJean-Francis RoyPublished in: CoRR (2015)
Keyphrases
- risk bounds
- bayesian analysis
- learning algorithm
- majority vote
- sample complexity
- vc dimension
- classifier ensemble
- computationally feasible
- machine learning algorithms
- fusion methods
- classifier combination
- active learning
- data dependent
- learning rate
- machine learning
- training data
- random selection
- reinforcement learning
- generalization ability
- learning problems
- supervised learning
- learning tasks
- statistical learning theory
- upper bound
- classification algorithm
- training samples
- learning process
- generalization error
- sample size
- labeled data
- pairwise
- pattern recognition
- objective function