Opinion formation in social systems: a network model-based simulation study

• Main Author: Mok, Wen Hao
• Other Authors: Xiao Gaoxi
• Format: Final Year Project
• Language: English
• Published: Nanyang Technological University 2024
• Subjects: Engineering; Complex network; Opinion evolution; Discrete/continuous value; Opinion evolution network model; Majority rule; Individual effect; Social opinion evolution; Deffuant-Weisbuch model; Bounded-confidence model; Consensus making; Social phenomena
• Online Access: https://hdl.handle.net/10356/176580
• Institution: Nanyang Technological University

Our opinions are formed through a mixture of personal experiences and the stream of information we encounter daily. However, our opinions are not immutable constructs. The interaction we hold with others as well as external factors can cause our opinions on certain topics to change. Opinion evolution has been studied in political polarization, cultural integration, the spread of rumours, and marketing strategies in social networks. Opinion evolution is about the changes in opinion state while network growth refers to how the structure of a network evolves over time. This differs based on different types of networks whether be it random network or scale-free network. Interaction patterns in networks are often studied in the field of opinion dynamics because it is intriguing how changing various conditions can drastically affect the opinion dynamics of various network models. In this dissertation, we will first be exploring various network models to have a better understanding of the different network topologies. Following which, we will be analysing the bounded confidence opinion formation model, the Deffuant-Weisbuch model and adding extensions to it to observe how such extensions would affect the opinion distribution. By altering the interactions agents have with one another, we are able to create simulations of varying versions of the Deffuant-Weisbuch model to simulate real-life social phenomenon and provide us with a deeper understanding of opinion dynamics in intricate social systems. We have utilized python programming to meticulously craft the various simulation extensions in our study. For our simulations, we first created a scenario where agents are only allowed to interact with one another if and only if both agents do not move out of the tolerance range of their current neighbours. Next, to further observe how the majority rule can affect our social system, we change the condition such that agents are allowed to interact as they do not decrease in number of neighbours within their tolerance range. Lastly, we also touched on the impact of zealots and how they can affect the overall opinion dynamics of our social system.