Darkpoppy / AffinityClustering

Hierarchical Graph clustering on sparse graphs using the Affinity Clustering algorithm of Bateni et al. (NIPS 2017)

Affinity Clustering

This repository provides the implemention of the Affinity Clustering method for partitioning sparse graphs introduced by Bateni et al. 2017. Apache Spark is used for performing the parallel computation.

The algorithm, along with testing code is provided in the following Jupyter Notebook.

Requirements

• Python 3.9
• PySpark 3.0.1
• numpy1.20.1
• networkx 2.5
• matplotlib 3.3.4
• sklearn
• scipy 1.6.1
• jupyter

Usage

Install the required dependencies

pip install -r requirements.txt

Recomended: use a dedicated virtual env or anaconda for easy dependency management