HiGNN: A Hierarchical Informative Graph Neural Network for Molecular Property Prediction Equipped with Feature-Wise Attention.
Weimin ZhuYi ZhangDuancheng ZhaoJianrong XuLing WangPublished in: J. Chem. Inf. Model. (2023)
Keyphrases
- neural network
- prediction model
- protein function prediction
- prediction accuracy
- graph structure
- random walk
- graph properties
- pattern recognition
- artificial neural networks
- elman network
- prediction error
- graph representation
- image features
- graph mining
- network model
- hierarchical structure
- directed graph
- vector representation
- graph theory
- multi layer perceptron
- three dimensional
- graph databases
- directed acyclic graph
- graph matching
- recurrent neural networks
- neural network model
- neural network ensemble
- structured data
- back propagation
- feature set
- feature vectors
- pairwise
- neural nets
- drug design
- anti monotonic
- global consistency
- feature maps
- graph theoretic
- spanning tree
- multi layer
- bipartite graph
- bp neural network
- fault diagnosis
- similarity measure
- genetic algorithm