Masters Topic: Modeling the Impact of Recommendation Diversity on Misinformation Spread in Social Networks: An Agent-based Approach

Defense Date: June 2025

Supervisors: Özlem Özgöbek, Önder Gürcan

Context

Social media platforms increasingly shape how people access news, but their recommendation algorithms may amplify misinformation and reinforce echo chambers by repeatedly presenting content aligned with users’ existing preferences. Although misinformation detection has been widely studied, the influence of recommendation strategies and content diversity on its propagation remains insufficiently understood.

Objective

This study aims to investigate how different recommendation algorithms affect misinformation spread and echo-chamber formation in social networks. Using an agent-based simulation, it compares popularity-based, collaborative-filtering, content-based, and random recommendations, and examines whether increasing recommendation diversity through Maximal Marginal Relevance can mitigate these effects.

Report

• Lise Jakobsen and Anna Johanne Holden Jacobsen. Modeling the Impact of Recommendation Diversity on Misinformation Spread in Social Networks: An Agent-based Approach [Master Report]. NTNU, Norway. June 2025.