University of Twente Student Theses

Login
This website will be unavailable due to maintenance December 1st between 8:00 and 12:00 CET.

Network research : exploration of centrality measures and network flows using simulation studies

Luo, Zhipeng (2018) Network research : exploration of centrality measures and network flows using simulation studies.

[img] PDF
2MB
Abstract:The study of network centrality has many real-world applications (e.g. targeted marketing strategy in business, identification of vital players in a criminal network, identification of keystone proteins in bio-research.) According to \citet{BORGATTI2005}, networks are differentiated by the flow mechanism (i.e. Parallel replication, Serial replication or Transfer) and the flow trajectory (i.e. Geodesic, Paths, Trails or Walks). However, the classification of networks showed that Freemans centrality measures are not applicable to various types of networks and flows. The aim of this research is to determine centrality measures that are applicable to varying network types and flows. An in-depth literature research was conducted to gain a deep understanding of various networks in terms of structure and information flow within the networks. In the process of deriving plausible centrality measures to address networks other than geodesic transfer network, betweenness-like (Flow centrality, Random-walk betweenness), diffusion, key players set and closeness-like measures were considered. The derived centrality approaches were validated using Entropy measures (Connectivity and Centrality Entropy). A simulation study was conducted to determine the impact of network properties (size, power-law/linkage parameter, clustering coefficient)(random vs. scale-free network) on all the proposed centrality measures. The results of different experiments showed that network cohesion is the main centrality factor. Network cohesion determines whether a network has a single central node or a set of key nodes. Furthermore, the network cohesion factor also proved that certain networks are comparable to some of the proposed centrality measures under certain network settings. However, the experiments also showed that the randomness factor had some impact on the results. Future research direction is aimed to address some of the limitations found in this research.
Item Type:Essay (Master)
Clients:
Unknown organization, Enschede, Netherlands
Faculty:EEMCS: Electrical Engineering, Mathematics and Computer Science
Subject:54 computer science
Programme:Business Information Technology MSc (60025)
Link to this item:https://purl.utwente.nl/essays/76847
Export this item as:BibTeX
EndNote
HTML Citation
Reference Manager

 

Repository Staff Only: item control page