A Graph Labelling Approach for Connected Feature Selection.
Jocelyn MarchadierSylvain MichelinYves EgelsPublished in: SSPR/SPR (2000)
Keyphrases
- feature selection
- connected components
- graph classification
- neighborhood graph
- graph theory
- text categorization
- feature space
- weighted graph
- structured data
- classification accuracy
- knn
- multi class
- graph structure
- mutual information
- bipartite graph
- graph model
- graph partitioning
- graph representation
- graph mining
- directed acyclic graph
- high dimensionality
- graph construction
- connected graphs
- k nearest neighbor
- feature set
- support vector machine
- feature extraction
- information gain
- graph matching
- feature subset
- data mining
- random walk
- text classification
- support vector
- method for feature selection
- machine learning