A Distributed Bagging Ensemble Methodology for Community Prediction in Social Networks

Christos Makris,Ioannis Orestis Rizos,Georgios Pispirigos

doi:10.3390/info11040199

Abstract

Presently, due to the extended availability of gigantic information networks and the beneficial application of graph analysis in various scientific fields, the necessity for efficient and highly scalable community detection algorithms has never been more essential. Despite the significant amount of published research, the existing methods—such as the Girvan–Newman, random-walk edge betweenness, vertex centrality, InfoMap, spectral clustering, etc.—have virtually been proven incapable of handling real-life social graphs due to the intrinsic computational restrictions that lead to mediocre performance and poor scalability. The purpose of this article is to introduce a novel, distributed community detection methodology which in accordance with the community prediction concept, leverages the reduced complexity and the decreased variance of the bagging ensemble methods, to unveil the subjacent community hierarchy. The proposed approach has been thoroughly tested, meticulously compared against different classic community detection algorithms, and practically proven exceptionally scalable, eminently efficient, and promisingly accurate in unfolding the underlying community structure.

Highlights

It is an indisputable fact that in 2019 more than 4.5 billion people used the internet on daily basis, while more than 2.7 billion of which were practically considered active social media users [1]
From methods that are exclusively based on the repetitive calculation of a global network topology metric, to alternatives inspired by discrete mathematics and physics, the pluralism of classic community detection processes is remarkable
Considering that the plain topological information of an edge might not be adequate for the proper extraction of the subjacent community information of an edge might not be adequate for the proper extraction of the subjacent community hierarchy, it is highly expected that the correct inter-connection/intra-connection classification might hierarchy, it is highly expected that the correct inter-connection/intra-connection classification might be significantly reliant to its kth-depth network topology information

Summary

Introduction

It is an indisputable fact that in 2019 more than 4.5 billion people used the internet on daily basis, while more than 2.7 billion of which were practically considered active social media users [1]. Graph analysis has been under the scientific spotlight, aiming to introduce efficient information analysis techniques but predominantly to define effective data mining methods. The iterative processes aiming to identify and remove all the inter-connection edges, by recursively maximizing a global topology criterion are classified as divisive algorithms [2,3,6,7,8]. In these techniques, at each iteration step, a finer community hierarchy layer is formed. The community detection evolution of any divisive algorithm could obviously be represented by dendrograms

Objectives

Methods

Results

Conclusion