Fig. 1: Overview of experimental setups.
From: Distinguishing mechanisms of social contagion from local network view
a Illustration of the three contagion mechanisms that are subject to inference: simple contagion parameterised by the adoption probability βi through a single stimuli; complex contagion parameterised by the threshold ϕi of necessary fraction of adopter neighbours to induce the adoption event; and spontaneous adoption that occurs with probability r. b The parameter space (β, ϕ) and the speed dependence of the simple and complex contagion processes are shown as a schematic representation for illustration purposes. c The different experimental setups that include the considered contagion mechanisms, the complexity of the underlying network, and model update rules. d Schematic pipeline for the application of the log-likelihood (LLH) and random forest machine learning (ML) classification approaches to the different experiments.
