Cluster homogeneity as a semi-supervised principle for feature selection using mutual information.
Frederico CoelhoAntônio de Pádua BragaMichel VerleysenPublished in: ESANN (2012)
Keyphrases
- mutual information
- feature selection
- semi supervised
- information theoretic
- conditional mutual information
- constrained clustering
- semi supervised clustering
- semi supervised learning
- information gain
- image registration
- clustering algorithm
- similarity measure
- labeled data
- unlabeled data
- semi supervised classification
- active learning
- medical image registration
- information theoretic measures
- supervised learning
- multi view
- pair wise constraints
- multimodal image registration
- data clustering
- multi class
- dimensionality reduction
- feature set
- text categorization
- support vector
- data points
- selection criterion
- hierarchical clustering
- machine learning
- feature subset
- clustering solutions
- classification accuracy
- clustering method
- model selection
- cluster analysis
- text classification
- subspace learning
- selected features
- co training
- feature selection algorithms
- unsupervised clustering
- duplicate detection
- irrelevant features
- normalized mutual information
- k means
- image analysis
- unsupervised learning