Finding Core-periphery Structure in Directed Networks

Finding Core-periphery Structure in Directed Networks: An Algorithm for Detecting Multiple-group Core-periphery Structure in Directed Netwoks

Huang, Hao (TU Delft Electrical Engineering, Mathematics and Computer Science; TU Delft Network Architectures and Services)

Kooij, R.E. (mentor)
Dubbeldam, J.L.A. (graduation committee)

Delft University of Technology

2021-08-31

The core-periphery structure is a mesoscale topological structure that refers to the presence of a dense core and a sparse periphery. The core-periphery structure has been discovered in financial, biological, and technological networks. Various methods for detecting core-periphery structure have been proposed, exploring different discrete or continuous models and single or multiple core-periphery groups. A method that implements multiple-group detection and edge-direction dependency to core-periphery detection is yet to be researched. This report proposes an algorithm to extract the core-periphery structure that satisfies multiple-group and edge-direction dependency requirements. This algorithm features a heuristic process and can process a large-scale network in an acceptable amount of time. Details of the theory behind the proposed algorithm are presented. The algorithm is tested on synthetic, random scale-free, and sampled dark web networks to verify the basic and advanced feasibility. Finally, in-depth analysis with the knowledge of core-periphery structure on a large-scale dark web network sample is presented.

Graph theory
Heuristic
Core-periphery Structure
Directed graph

http://resolver.tudelft.nl/uuid:7b7ab297-e97f-4c42-9f4f-48b694172e09

Student theses

master thesis

© 2021 Hao Huang