Rank-based self-training for graph convolutional networks.
Daniel Carlos Guimarães PedronetteLongin Jan LateckiPublished in: Inf. Process. Manag. (2021)
Keyphrases
- average degree
- dynamic networks
- highly connected
- graph structure
- semi supervised classification
- directed edges
- random walk
- betweenness centrality
- directed graph
- graph theory
- fully connected
- graph representation
- social networks
- community discovery
- small world
- graph theoretic
- graph model
- co training
- neural network
- graph layout
- shortest path
- network structure
- social graphs
- complex networks
- graph structures
- scale free
- deep learning
- connected components
- overlapping communities
- real world graphs
- sparse coding
- network analysis
- convolutional neural networks
- link formation
- path length