An agent-based simulation study of the social contagion of divorce

Abstract

An agent-base model based on an epidemiological SIRa model of infectious diseases is used for the simulation of the social contagion of divorce in this work. First, two simple SIR and SIRa models are implemented and the results of their simulation outcomes analyzed. Then, using machine learning classification models to predict the probability that each couple has of facing relationship instability, multiple modifications are introduced in the SIRa model. Having more realistic and heterogeneous agents in the agent-based simulation leads to more realistic and accurate results and less uncertainty in the outcomes of the simulation runs.