Single-Class Classification Augmented with Unlabeled Data: A Symbolic Approach.
Andrew SkabarPublished in: Australian Conference on Artificial Intelligence (2003)
Keyphrases
- unlabeled data
- single class
- labeled data
- semi supervised learning
- class labels
- supervised learning
- co training
- semi supervised
- text classification
- training set
- multi class
- active learning
- training data
- learning algorithm
- ensemble learning
- classification error
- classification accuracy
- label information
- decision boundary
- training examples
- binary classification
- labeled examples
- machine learning
- text categorization
- classification algorithm
- class distribution
- transfer learning
- feature selection
- pattern classification
- decision trees
- positive data
- prior knowledge
- binary classifiers
- image classification
- data points
- positive examples
- pairwise
- data sets
- class probabilities
- feature extraction
- reinforcement learning
- machine learning algorithms
- feature vectors
- unsupervised learning
- learning problems
- positive and negative
- naive bayes