Feature selection approach based on hypothesis-margin and pairwise constraints.
Samah HijaziMariam KalakechDenis HamadAli KalakechPublished in: MENACOMM (2018)
Keyphrases
- pairwise constraints
- feature selection
- decision boundary
- semi supervised
- support vector
- pairwise
- support vector machine
- semi supervised clustering
- loss function
- boosting algorithms
- spectral clustering
- labeled data
- information gain
- text categorization
- multi class
- hyperplane
- pattern classification
- document clustering
- unlabeled data
- data points
- training set
- data representation
- unsupervised learning
- maximum margin
- machine learning
- semi supervised learning
- text classification
- feature set
- classification accuracy
- feature extraction
- support vector machine svm
- model selection
- dimensionality reduction
- constrained clustering
- feature subset
- nearest neighbor
- linear classifiers
- knn
- multi task
- information retrieval
- data sets