Properties and Performance of the ABCDe Random Graph Model with Community Structure

TL;DR

This paper introduces ABCDe, a highly efficient multi-threaded random graph generator with community structure, outperforming existing models in speed and scalability while maintaining comparable graph properties.

Contribution

We present ABCDe, a multi-threaded implementation of the ABCD model, which is faster and scales better than existing LFR generators, with similar graph properties.

Findings

ABCDe is over ten times faster than parallel LFR in NetworKit.

ABCDe scales better with larger graphs.

Generated graphs have properties similar to those from the original LFR model.

Abstract

In this paper, we investigate properties and performance of synthetic random graph models with a built-in community structure. Such models are important for evaluating and tuning community detection algorithms that are unsupervised by nature. We propose ABCDe, a multi-threaded implementation of the ABCD (Artificial Benchmark for Community Detection) graph generator. We discuss the implementation details of the algorithm and compare it with both the previously available sequential version of the ABCD model and with the parallel implementation of the standard and extensively used LFR (Lancichinetti--Fortunato--Radicchi) generator. We show that ABCDe is more than ten times faster and scales better than the parallel implementation of LFR provided in NetworKit. Moreover, the algorithm is not only faster but random graphs generated by ABCD have similar properties to the ones generated by the original LFR algorithm, while the parallelized NetworKit implementation of LFR produces graphs that have noticeably different characteristics.