AutoEval: Are Labels Always Necessary for Classifier Accuracy Evaluation?
Weijian DengLiang ZhengPublished in: IEEE Trans. Pattern Anal. Mach. Intell. (2024)
Keyphrases
- training data
- class labels
- unseen data
- training set
- labeled instances
- fold cross validation
- labeling effort
- high accuracy
- training examples
- prediction accuracy
- classification rate
- classification accuracy
- roc curve
- associative classifiers
- multilabel classification
- highest accuracy
- fully supervised
- support vector
- significantly improves the accuracy
- computational cost
- active learning
- ground truth labels
- confusion matrix
- annotation effort
- improving classification accuracy
- semantic labels
- support vector machine classifier
- decision trees
- leave one out cross validation
- class membership
- feature reduction
- multiple classifier systems
- individual classifiers
- classifier combination
- multi label classification
- classification scheme
- training samples
- labeled data
- feature set
- feature space