Bipartite graph analysis as an alternative to reveal clusterization in complex systems

TL;DR

This paper proposes using bipartite graph co-clustering to analyze community structures in complex systems, specifically applied to scientific knowledge networks, enabling inference of concept features from article clusters.

Contribution

It introduces a bipartite co-clustering approach as an alternative method to reveal and compare community structures in bipartite networks, bridging different clustering results.

Findings

Effective connection of article and concept clusters.

Inference of concept features from article data.

Application to scientific knowledge networks.

Abstract

We demonstrate how analysis of co-clustering in bipartite networks may be used as a bridge to connect, compare and complement clustering results about community structure in two different spaces: single-mode bipartite network projections. As a case study we consider scientific knowledge, which is represented as a complex bipartite network of articles and related concepts. Connecting clusters of articles and clusters of concepts via article-to-concept bipartite co-clustering, we demonstrate how concept features (e.g. subject classes) may be inferred from the article ones.