INK: knowledge graph embeddings for node classification.
Bram SteenwinckelGilles VandewieleMichael WeynsTerencio AgozzinoFilip De TurckFemke OngenaePublished in: Data Min. Knowl. Discov. (2022)
Keyphrases
- graph structure
- feature vectors
- classification scheme
- knowledge base
- decision trees
- graph theory
- image classification
- pattern recognition
- directed graph
- classification accuracy
- support vector machine svm
- machine learning
- text classification
- classification method
- knowledge discovery
- expert systems
- support vector
- bipartite graph
- overlapping communities
- knowledge acquisition
- prior knowledge
- feature space
- bayesian networks
- feature extraction
- random walk
- domain experts
- tree structure
- classification algorithm
- distance measure
- classification rules
- connected components
- knowledge representation
- manifold learning
- graph databases
- graph representation
- feature selection