Dual-Attention Multi-Scale Graph Convolutional Networks for Highway Accident Delay Time Prediction.
I-Ying WuFandel LinHsun-Ping HsiehPublished in: SIGSPATIAL/GIS (2021)
Keyphrases
- multiscale
- prediction accuracy
- traffic accidents
- average degree
- dynamic networks
- graph structures
- edge weights
- prediction model
- social networks
- graph layout
- highly connected
- neural network ensemble
- overlapping communities
- graph theoretic
- graph representation
- graph structure
- directed graph
- natural images
- protein function prediction
- traffic safety
- network structure
- prediction error
- bipartite graph
- weighted graph
- connected components
- directed acyclic graph
- scale space
- betweenness centrality
- citation networks
- community discovery
- fully connected
- random walk
- image representation
- small world
- multiple scales
- primal dual
- graph partitioning
- visual attention
- network analysis
- road safety
- social graphs
- bayesian networks
- image processing
- neural network
- directed edges
- edge detection
- degree distribution
- random graphs
- power law
- structured data
- graphical representation
- complex networks
- graph databases