Designing Labeled Graph Classifiers by Exploiting the Rényi Entropy of the Dissimilarity Representation.
Lorenzo LiviPublished in: Entropy (2017)
Keyphrases
- dissimilarity representation
- structural pattern recognition
- training data
- pattern recognition
- training set
- shannon entropy
- labeled graphs
- graph theory
- test set
- labeled training data
- graph representation
- mutual information
- random walk
- svm classifier
- statistical pattern recognition
- classification algorithm
- dissimilarity measure
- information theory
- structured data
- decision trees
- information theoretic
- directed graph
- feature set
- graph structure
- support vector
- graph mining
- neural network
- feature selection
- feature extraction
- unlabeled samples
- supervised learning
- bipartite graph
- training examples
- nearest neighbor
- training instances
- weighted graph
- graph matching
- learning algorithm
- machine learning