Cost-sensitive multi-label learning with positive and negative label pairwise correlations.
Guoqiang WuYingjie TianDalian LiuPublished in: Neural Networks (2018)
Keyphrases
- positive and negative
- cost sensitive
- multi label learning
- pairwise
- multi label
- multi class
- base learners
- semi supervised
- binary classification
- multi label classification
- multiple labels
- misclassification costs
- text categorization
- label propagation
- active learning
- labeled data
- decision rules
- image classification
- markov random field
- class distribution
- graph cuts
- multi instance
- similarity measure
- unlabeled data
- loss function
- pairwise constraints
- semi supervised learning
- text classification
- image annotation
- higher order
- learning tasks
- class imbalance
- multi class classification
- image processing
- support vector machine
- data sets
- neural network
- text mining
- machine learning methods
- decision trees
- feature selection
- boosting algorithms
- computer vision
- naive bayes
- machine learning