Opinion Polarization — Echo Chamber Model

Bounded confidence + algorithmic homophily drive opinion clusters

Agents hold opinions in [0,1] and interact only when their opinion difference is within tolerance ε (Deffuant-Weisbuch bounded confidence). Algorithmic homophily μ controls how strongly the network preferentially connects similar agents, mimicking social media recommendation. Varying ε and μ produces consensus, polarization, or fragmentation.