https://github.com/giannisnik/cnn-graph-classification

A convolutional neural network for graph classification in PyTorch
https://github.com/giannisnik/cnn-graph-classification

convolutional-neural-networks graph-classification graph-kernels

Last synced: 22 days ago
JSON representation

A convolutional neural network for graph classification in PyTorch

• Host: GitHub
• URL: https://github.com/giannisnik/cnn-graph-classification
• Owner: giannisnik
• Created: 2017-08-17T15:41:57.000Z (over 7 years ago)
• Default Branch: master
• Last Pushed: 2019-02-15T17:11:49.000Z (almost 6 years ago)
• Last Synced: 2024-08-01T03:46:07.285Z (4 months ago)
• Topics: convolutional-neural-networks, graph-classification, graph-kernels
• Language: Python
• Homepage:
• Size: 3.09 MB
• Stars: 90
• Watchers: 7
• Forks: 20
• Open Issues: 4
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

• awesome-graph-classification - [Python Reference

README

        
## Kernel Graph Convolutional Neural Networks

Code for the paper [Kernel Graph Convolutional Neural Networks](https://link.springer.com/chapter/10.1007/978-3-030-01418-6_3).

### Requirements

Code is written in Python 3.6 and requires:

* PyTorch 0.3

* NetworkX 1.11

* igraph 0.7

* scikit-learn 0.18

### Datasets

Use the following link to download datasets: 

```

https://ls11-www.cs.tu-dortmund.de/staff/morris/graphkerneldatasets

```

Extract the datasets into the `datasets` folder.

### Run the model

First, specify the dataset and the hyperparameters in the `main.py` file. Then, use the following command:

```

$ python main.py

```

### Cite

Please cite our paper if you use this code:

```

@inproceedings{nikolentzos2018kernel,

  title={Kernel Graph Convolutional Neural Networks},

  author={Nikolentzos, Giannis and Meladianos, Polykarpos and Tixier, Antoine Jean-Pierre and Skianis, Konstantinos and Vazirgiannis, Michalis},

  booktitle={International Conference on Artificial Neural Networks},

  pages={22--32},

  year={2018},

  organization={Springer}

}

```

-----------

Provided for academic use only