Post-Training Overfitting Mitigation in DNN Classifiers.
Hang WangDavid J. MillerGeorge KesidisPublished in: CoRR (2023)
Keyphrases
- training process
- training set
- training data
- test set
- training examples
- training samples
- decision trees
- avoid overfitting
- adaboost algorithm
- cross validation
- unseen data
- trained classifiers
- supervised learning
- linear svm
- machine learning methods
- classifier training
- training algorithm
- feature selection
- supervised classification
- classification rate
- active learning
- training set size
- machine learning algorithms
- training and testing data
- training error
- supervised training
- training stage
- labeled training data
- feature selection and classifier
- training phase
- linear classifiers
- generalization error
- classification accuracy
- support vector
- text classifiers
- object detectors
- multi layer perceptron
- extracted features
- roc curve
- ensemble learning
- risk management
- feature set
- learning algorithm