Mesoscopic analysis of networks: applications to exploratory analysis and data clustering

TL;DR

This paper adapts modularity-based algorithms from complex network analysis to explore mesoscopic structures in correlation matrices, enabling multi-resolution clustering for applications like neural connectivity analysis and data classification.

Contribution

It introduces a multi-resolution approach to analyze correlation matrices using network modularity algorithms, extending their application to data clustering and neural connectivity.

Findings

Successful analysis of C. elegans neural connectivity

Effective classification of Iris dataset

Demonstrated multi-resolution clustering capability

Abstract

We investigate the adaptation and performance of modularity-based algorithms, designed in the scope of complex networks, to analyze the mesoscopic structure of correlation matrices. Using a multi-resolution analysis we are able to describe the structure of the data in terms of clusters at different topological levels. We demonstrate the applicability of our findings in two different scenarios: to analyze the neural connectivity of the nematode {\em Caenorhabditis elegans}, and to automatically classify a typical benchmark of unsupervised clustering, the Iris data set, with considerable success.