0.BN-GNN
Python implementation for "Deep reinforcement learning guided graph neural networks for brain network analysis"

1.Overview
data.tar: Compression of three public datasets.
/model/dqn.py: Realize the meta-policy module in BN-GNN.
/model/gnn.py: Realize the network building module and the GNN module in BN-GNN.
main.py: Entry file of BN-GNN.

2.Example
python main.py

3.Requirements
python=3.6.7
pytorch=1.4.0
cudnn=7.6.5
numpy=1.19.5
scipy=1.2.1
deepwalk=1.0.3

4.Baselines
DeepWalk: "Deepwalk: Online learning of social representations"
Node2Vec: "Node2vec: Scalable feature learning for networks"
GCN: "Semi-supervised classification with graph convolutional networks"
GAT: "Graph attention networks"
GraphSAGE: "Inductive representation learning on large graphs"
FastGCN: "FastGCN: Fast learning with graph convolutional networks via importance sampling"
PR-GNN: "Pooling regularized graph neural network for fmri biomarker analysis"
GNEA: "GNEA: A graph neural network with ELM aggregator for brain network classification"
Hi-GCN: "Hi-GCN: A hierarchical graph convolution network for graph embedding learning of brain network and brain disorders prediction"
SDBN: "Structural deep brain network mining"

5.Public Datasets
ADHD-fMRI & HI-fMRI & GD-fMRI
"Task sensitive feature exploration and learning for multitask graph classification"
"A whole brain fMRI atlas generated via spatially constrained spectral clustering"