Statistical physics of opinion formation and election

Abstract

A society may be treated as a many-particle system so that opinion formation therein may be modelled and analysed with statistical physics, then winners and losers in elections may be arguably predicted. Different models capture different aspects of voters’ psychological behaviours, yet they all fit into the framework introduced in Chapter 1. Subsequently, Chapter 2 and Chapter 3 particularly study the Ising-Glauber model and the exponential voter model. The former suggests that a key opinion leader’s degree on a social network best indicates his influence on the society. The latter, which is mathematically simpler and psychologically more intuitive, can yield various analytical solutions that describe the evolution of a society’s public opinion. Chapter 4 proceeds to investigate numerical properties of different multi-party models generalized from the two models. As Chapter 5 concludes, despite their distinct formulations, all the models universally hint that voters’ engagement in communication and thus the resultant dimensionality of a social network may crucially affect the extent of consensus and the rate of its emergence.

Date of Award	2022
Original language	English
Awarding Institution	The Hong Kong University of Science and Technology
Supervisor	Bradley Alan FOREMAN (Supervisor), Kwok Yee Michael WONG (Supervisor) & Sai Ping LI (Supervisor)