Author Date

2020-06-12

Degree Name

Department

Mathematics

College

Physical and Mathematical Sciences

Defense Date

2020-05-29

Publication Date

2020-06-14

First Faculty Advisor

Benjamin Webb

Second Faculty Advisor

Emily Evans

First Faculty Reader

Benjamin Webb

Second Faculty Reader

Emily Evans

Honors Coordinator

Michael Griffin

Keywords

Networks, Social Networks, Machine Learning, Community Detection

Abstract

A well-studied topic in network theory is detecting the communities found in real-world networks. Community detection is a technique to better understand the way in which small dense substructures appear in these networks. Such substructures can often tell important information about groups that form in such systems. A prominent feature of many networks is that they evolve over time, forming and dissolving new edges between different nodes that appear. In this thesis, we consider how we can use the community structure of a network at a certain point in time to predict the state of a network’s communities at some time in the future. Through the use of ”affinity scores” that describe a node’s inclination to be part of a community, we predict the formation of future communities in the network with the assistance of machine learning algorithms. Using the method proposed in this thesis, we find that it is possible to predict, with a moderate degree of accuracy, the communities that will eventually form in a network before they are fully formed.