Social Network Analysis and Mining manuscript No. (will be inserted by the editor) Development and Evaluation of Multi-Agent Models of Online Social Influence based on Cialdini’s Principles

Abstract The aim of this study is to better understand social influence in online social media. Therefore, we propose a method in which we implement, validate and improve individual behavior models. The behavior model is based on three fundamental behavioral principles of social influence from the liter-ature (i.e., Cialdini’s principles): 1) consistency, 2) liking and 3) social proof. We have implemented the model using an agent-based modeling approach. The multi-agent model contains the social network structure, individual be-havior parameters and the scenario that are obtained from empirical data. The model is validated by comparing the output of the multi-agent simulation with empirical data. We demonstrate the method by evaluating five versions of behavior models applied to the use case of Twitter behavior about a talent show on Dutch television