Fundamental Limits of Deep Learning-Based Binary Classifiers Trained with Hinge Loss.
Tilahun Melkamu GetuGeorges KaddoumPublished in: CoRR (2023)
Keyphrases
- binary classifiers
- binary classification
- multi class
- svm classifier
- support vector machine
- loss function
- support vector
- cross validation
- cost sensitive
- training set
- learning problems
- hyperplane
- training samples
- multi label
- support vector machine svm
- pairwise
- ensemble methods
- unsupervised learning
- generalization error
- data sets
- kernel function
- image classification
- high resolution
- learning process
- feature vectors
- input space
- feature space
- training data
- feature selection
- machine learning