Cost-Sensitive Multi-Label Feature Selection Algorithm Based on Positive Approximation.
Binglong WuWenbin QianQin HuangSudan YuPublished in: FSDM (2018)
Keyphrases
- multi label
- cost sensitive
- feature selection algorithms
- feature selection
- text categorization
- binary classification
- misclassification costs
- text classification
- graph cuts
- cost sensitive learning
- data sets
- image classification
- learning tasks
- class distribution
- naive bayes
- class imbalance
- multi instance
- feature subset
- feature set
- multi class
- learning models
- classification accuracy
- feature extraction
- knn
- support vector
- machine learning
- class labels
- small number
- base classifiers
- information retrieval