Novel tight classification error bounds under mismatch conditions based on f-Divergence.
Ralf SchlüterMarkus Nußbaum-ThomEugen BeckTamer AlkhouliHermann NeyPublished in: ITW (2013)
Keyphrases
- error bounds
- worst case
- classification algorithm
- lower bound
- pattern recognition
- machine learning
- automatic classification
- feature extraction
- theoretical analysis
- image classification
- classification systems
- sufficient conditions
- feature vectors
- support vector
- upper bound
- support vector machine
- classification method
- approximation algorithms
- machine learning methods
- generalization error bounds
- feature selection
- supervised classification
- classification scheme
- face recognition
- pattern classification
- training data
- training samples
- classification accuracy
- neural network
- denoising