This is an implement of the DGLC method in "Jinyu Cai, Yi Han, Wenzhong Guo, and Jicong Fan, Deep graph-level clustering using pseudo-label-guided mutual information maximization network, Neural Computing and Applications, 2024".

• You can run the model with some hyperparameters. For example:

  python DGLC.py --DS BZR --lr 0.00001 --num-gc-layers 4 --hidden-dim 32 --cluster_emb 25
  

• The supported database is TUDataset, which will be downloaded automatically at runtime (if it does not exist locally).

• DS is the dataset name. You can change from (MUTAG, PTC-MR, BZR, PTC-MM, ENZYMES, COX2), and the dataset will be downloaded automatically.
• num-gc-layers is the number of GNN hidden layers.
• hidden-dim is the dimension of the hidden layer of GNN.
• cluster_emb is the dimension of the cluster projector.

If you use the code or find this repository useful for your research, please consider citing our paper.

@article{cai2024dglc,
  title = {Deep graph-level clustering using pseudo-label-guided mutual information maximization network},
  author = {Cai, Jinyu and Han, Yi and Guo, Wenzhong and Fan, Jicong},
  journal = {Neural Computing and Applications},
  pages = {1--16},
  year = {2024},
  publisher = {Springer}}