Epidemic spreading Complex Networks 2020

This repository contains code for the paper: Prediction of the effects of epidemic spreading with graph neural networks from the Complex Networks 2020 conference. The paper can be found here and cited as:

@InProceedings{meznar2020spreading,
    author="Me{\v{z}}nar, Sebastian and Lavra{\v{c}}, Nada and {\v{S}}krlj, Bla{\v{z}}",
    editor="Benito, Rosa M. and Cherifi, Chantal and Cherifi, Hocine
        and Moro, Esteban and Rocha, Luis Mateus and Sales-Pardo, Marta",
    title="Prediction of the Effects of Epidemic Spreading with Graph Neural Networks",
    booktitle="Complex Networks {\&} Their Applications IX",
    year="2021",
    publisher="Springer International Publishing",
    address="Cham",
    pages="420--431",
    isbn="978-3-030-65347-7"
}

An overview of the approach can be seen in the image below.

Running the code

Intstall the required python packages using the command:

pip install -r requirements.txt

The evaluation of machine learning methods can be ran using the command:

python evaluate_epidemics.py  

Simulations from the original paper are marked with name old. The results of these simulations is approximately shown in the two tables below.

Interpreting CABoost model

The prediction can be interpreted using tools such as SHAP. An example of this can be tested by running the script interpreting_effects.py as:

python interpreting_effects.py

Running the code gives a waterfall plot such as the one on the image below

Simulations

Additional simulation data can be created using the run create_data.py script. The format of simulation data is "{time} {score}" but should be changed to "{node} {time} {score}" for usage.