Detecting Changes in Opinion Value Distribution for Voter Model

We address the problem of detecting the change in opinion share over a social network caused by an unknown external situation change under the valueweighted voter model with multiple opinions in a retrospective setting. The unknown change is treated as a change in the value of an opinion which is a model parameter, and the problem is reduced to detecting this change and its magnitude from the observed opinion share diffusion data. We solved this problem by iteratively maximizing the likelihood of generating the observed opinion share, and in doing so we devised a very efficient search algorithm which avoids parameter value optimization during the search. We tested the performance using the structures of four real world networks and confirmed that the algorithm can efficiently identify the change and outperforms the naive method, in which an exhaustive search is deployed, both in terms of accuracy and computation time.