Localized Densest SubHypergraph Problem

This repo contains codes for the following paper:

Densest Subhypergraph: Negative Supermodular Functions and Strongly Localized Methods

If you feel it helpful, please cite the paper mentioned above.

Environment

We use julia1.9 and we provide the Project.toml and Manifest.toml files for our environment. If you are using REPL, you should be able to activate it by typing activate . in Pkg REPL(i.e. ] activate .). If you are running the code files from the command line, remember to include --project(i.e. julia --project filename.jl).

Download Data and Initialization

The five real-world datasets we use can be accessed at Hypergraph Datasets. For convenience, run cd src; bash init.sh to create all directories and download the datasets.

The procedures of generating synthetic data and accessing the web graph data are included in the corresponding directories.

Basic Usage

./src contains all the codes, experiment-specific codes are included in ./src/Exp... folders, others are commonly used codes shared through experiments.

You can use all our interfaces via

include("headers.jl")

and the documentation is available via typing ? and the function name in REPL.

The README for each experiment is included in each subfolder.

Code Structure

In general, we maintain the following code structure:

|_src/
|_data/
      |_datasetname/
|_figs/
      |_datasetname/
|_results/
      |_datasetname/

The data for each dataset is saved in the ./data/datasetname/ folder, the running results of it are saved in the ./results/datasetname/ folder, and the plotted figures are saved in the ./figs/datasetname/. You may need to create those folders before experiments and modify savepath in the codes.

Acknowledgement

Part of this code is inspired by AnchoredDensestSubgraph and HypergraphFlowClustering.