An agent-based modelling approach to analyse the public opinion on politicians

Abstract

A politician’s popularity can be measured by polls or by measuring the amount of times a politician is mentioned on the Internet in a positive or negative manner. This paper introduces an approach to an agent-based computational model to model a politician’s popularity within a population that participates on the Internet over time. A particle swarm optimization algorithm is used for parameter tuning to identify the characteristics of all agents based on the analysis of public opinions on a politician found on the Internet. The properties of the network are verified by applying a social network analysis. A mathematical analysis is used to get more in depth understanding on the model and to verify its correctness.