A Generative Neighborhood-Based Deep Autoencoder with an Extended Loss Function for Robust Imbalanced Classification.
Eirini TroullinouGrigorios TsagkatakisAttila LosonczyPanayiota PoiraziPanagiotis TsakalidesPublished in: ACSSC (2023)
Keyphrases
- loss function
- support vector
- hinge loss
- logistic regression
- pairwise
- classification accuracy
- bayes rule
- generative model
- feature extraction
- learning to rank
- support vector machine svm
- generalization ability
- unsupervised learning
- supervised learning
- class imbalance
- reproducing kernel hilbert space
- empirical risk
- machine learning
- decision trees
- boosting framework
- decision rules
- class labels
- benchmark datasets
- image classification
- support vector machine
- probabilistic model
- classification algorithm
- training samples
- text classification
- stochastic gradient descent
- risk minimization
- feature vectors
- feature space
- convex loss functions