Are Labels Necessary for Classifier Accuracy Evaluation?
Weijian DengLiang ZhengPublished in: CoRR (2020)
Keyphrases
- training data
- unseen data
- confusion matrix
- fold cross validation
- classification accuracy
- training set
- prediction accuracy
- classification rate
- high accuracy
- decision trees
- class labels
- labeling effort
- significantly improves the accuracy
- fully supervised
- higher classification accuracy
- leave one out cross validation
- pairwise
- multi label classification
- feature reduction
- face detection
- highest accuracy
- supervised learning
- high classification accuracy
- supervised classifiers
- feature selection
- annotation effort
- individual classifiers
- classifier combination
- classification scheme
- svm classifier
- error rate
- training examples
- training samples
- labeled data
- feature set
- computational cost
- active learning
- support vector