BoCluSt: Bootstrap clustering stability algorithm for community detection

5Citations
Citations of this article
18Readers
Mendeley users who have this article in their library.

Abstract

The identification of modules or communities in sets of related variables is a key step in the analysis and modeling of biological systems. Procedures for this identification are usually designed to allow fast analyses of very large datasets and may produce suboptimal results when these sets are of a small to moderate size. This article introduces BoCluSt, a new, somewhat more computationally intensive, community detection procedure that is based on combining a clustering algorithm with a measure of stability under bootstrap resampling. Both computer simulation and analyses of experimental data showed that BoCluSt can outperform current procedures in the identification of multiple modules in data sets with a moderate number of variables. In addition, the procedure provides users with a null distribution of results to evaluate the support for the existence of community structure in the data. BoCluSt takes individual measures for a set of variables as input, and may be a valuable and robust exploratory tool of network analysis, as it provides 1) an estimation of the best partition of variables into modules, 2) a measure of the support for the existence of modular structures, and 3) an overall description of the whole structure, which may reveal hierarchical modular situations, in which modules are composed of smaller sub-modules.

Cite

CITATION STYLE

APA

Garcia, C. (2016). BoCluSt: Bootstrap clustering stability algorithm for community detection. PLoS ONE, 11(6). https://doi.org/10.1371/journal.pone.0156576

Register to see more suggestions

Mendeley helps you to discover research relevant for your work.

Already have an account?

Save time finding and organizing research with Mendeley

Sign up for free