Home WakeSpace Scholarship › Electronic Theses and Dissertations

An Analysis of the Practicality of Network Centrality Measures in Evaluation of Real-World Characteristics on Social Networks

Electronic Theses and Dissertations

Item Files

Item Details

abstract
Network analysis is a burgeoning field, showing success at explaining various characteristics in social network contexts. Centrality is a key property used in analysis of real world networks. Many different measures for centrality are used, and a number of other measures have been proposed in non-centrality frameworks which may prove more related to real-world characteristics on human networks. These centrality measures are defined by informational flow across a network. Some evidence shows that measures such as well-being and time spent in a network may be predicted by centrality despite a lack of ``flow". This paper investigates four Wake Forest University Greek Letter Organizations, creating social network datasets from each one to determine whether well-being and time spent are correlated with centrality, and uses several common measures of centrality along with two newer analytical techniques - partitioned local depth (PaLD) and reciprocation values - to analyze what measures on the network best predict social well-being and time as member of the chapter. Analysis shows clustering-style measures such as assortativity are best for predicting time spent. Multiple measures combine to form the best prediction of social well-being, with one local depth measure via PaLD appearing to capture other trends in well-being more frequently and consistently than established centrality measures.
subject
Centrality
Friendship Networks
Network Analysis
Partitioned Local Depth
Social Networks
contributor
Tague, Mitchell Carlson (author)
Berenhaut, Kenneth (committee chair)
Jiang, Miaohua (committee member)
date
2020-08-28T08:35:24Z (accessioned)
2020 (issued)
degree
Mathematics and Statistics (discipline)
2025-09-01 (liftdate)
embargo
2025-09-01 (terms)
identifier
http://hdl.handle.net/10339/96945 (uri)
language
en (iso)
publisher
Wake Forest University
title
An Analysis of the Practicality of Network Centrality Measures in Evaluation of Real-World Characteristics on Social Networks
type
Thesis

Usage Statistics