Correlation Analysis of Nodes Identifies Real Communities in Networks

TL;DR

This paper introduces a simple, correlation-based algorithm for community detection in networks that does not require prior information or optimization, effectively identifying community structures in both known and unknown networks.

Contribution

The paper presents a novel, efficient community detection algorithm based solely on node correlation, avoiding the need for predefined parameters or multiple runs.

Findings

Accurately detects known community structures in synthetic and real-world networks.

Effectively identifies meaningful communities in networks with unknown structures.

Operates reliably without requiring prior information or multiple iterations.

Abstract

A significant problem in analysis of complex network is to reveal community structure, in which network nodes are tightly connected in the same communities, between which there are sparse connections. Previous algorithms for community detection in real-world networks have the shortcomings of high complexity or requiring for prior information such as the number or sizes of communities or are unable to obtain the same resulting partition in multiple runs. In this paper, we proposed a simple and effective algorithm that uses the correlation of nodes alone, which requires neither optimization of predefined objective function nor information about the number or sizes of communities. We test our algorithm on real-world and synthetic graphs whose community structure is already known and observe that the proposed algorithm detects this known structure with high applicability and reliability. We also apply the algorithm to some networks whose community structure is unknown and find that it detects deterministic and informative community partitions in these cases.