In this repository we propose a solution to the DOS problem via Agglomerative Greedy Enumeration (DOSAGE) algorithm as a novel approach to enhance the process of generating the densest overlapping subgraphs and, hence, a robust construction of the hypergraphs. Experiments on standard benchmarks show that the DOSAGE algorithm significantly outperforms the HGNNs and six other methods on the node classification task.

This is a code for Hyperedge Modeling in Hypergraph Neural Networks by using Densest Overlapping Subgraphs paper. You can read the paper in here: https://arxiv.org/abs/2409.10340

Follow these steps to set up the project locally.

• Python (>= 3.8)
• Conda (or Miniconda)

1. Clone the repository:

   git clone hhttps://github.com/Mehrads/Hypergraph-Modelling-using-DOSAGE.git
   cd Hypergraph-Modelling-using-DOSAGE

Create a virtual environment:

conda create --name env python=3.8

Activate the virtual environment:

conda activate env

Install dependecies:

conda install --file requirements.txt