POTION

Code to replicate experimental results in the paper.

1. Use conda to create a virtual environment named POTION:

conda create -n POTION python=3.7

2. Activate the virtual environment:

conda activate POTION

3. Install necessary packages (PyTorch version is 1.4.0):

conda install numpy scipy pandas matplotlib seaborn pandas networkx=2.4 jupyter
conda install pytorch torchvision -c pytorch
pip install EoN

4. Create a folder at the root directory to save results:

mkdir result/

5. Check out the ipython notebook src/exp.ipynb for an example of running the code. In order to run POTION on networks not listed in the paper, please check out the example at the end of the notebook.

Reference

@article{yu2020optimizing,
  title={Optimizing Graph Structure for Targeted Diffusion},
  author={Yu, Sixie and Torres, Leonardo and Alfeld, Scott and Eliassi-Rad, Tina and Vorobeychik, Yevgeniy},
  journal={arXiv preprint arXiv:2008.05589},
  year={2020}
}