GNN-proces-discovery

This repository contains the code developed for the scientific article "A Machine Learning-Based Process Mining Discovery Approach".
The repository provides the tools and scripts needed to reproduce the experiments and results presented in the article.

---

Table of Contents

• Introduction
• Installation
• Usage
• License
• Acknowledgements

---

Introduction

This project presents a novel algorithm for process discovery that leverages graph neural networks to infer sound Petri nets from event logs. The code enables users to replicate the experimental results and adapt the method for further research.

---

Installation

Clone the repository and install the requirements:

git clone https://github.com/jaxels20/GNN-proces-discovery.git
cd GNN-proces-discovery
python -m venv env
source env/bin/activate
pip install -r requirements.txt

Usage

1. Generate Synthetic Data

      python3 data_generation/data_generation.py 

2. Train Model

      python3 training.py 

3. Controlled Scenario Evaluation

      python3 evaluate_on_controlled_scenarios.py

4. Real Life Evaluation

      python3 evaluate_on_reallife_datasets.py 

License

This repository is licensed under the MIT License. See the LICENSE file for details.

Acknowledgements