Propositional Opinion Diffusion with Constraints
Sirin Botan

Abstract:
This thesis examines influence-based iterative opinion diffusion mechanisms on social networks where agents update their opinions based on the opinions of their influencers in the social network. We use tools from judgment aggregation to analyse the strengths and weaknesses of an existing opinion diffusion framework, Propositional Opinion Diffusion, and define a variant of this type of diffusion mechanism, Propositionwise Opinion Diffusion, which guarantees the rationality of agents’ opinions at every step of the diffusion process. We compare the two variants of the mechanism, both by using an axiomatic approach, and by exploring when they might coincide or differ on the opinions agents hold when the mechanisms terminate.