Formal models of opinion formation and their application to real data: evidence from online social networks

ABSTRACT

In this paper, we analyze data on the opinion dynamics of 1,660,927 users of an online social network using formal models of opinion formation. We have observed that moderate users have a tendency to follow the average opinion of their online friends, which we interpret as a presence of bounded confidence. Further, we have discovered that the probability of moving toward the average opinion goes down if the difference between it and the user’s opinion is too large. Another interesting feature uncovered is that if a user’s opinion and the average opinion of their online friends are very similar, the influence also decreases.

KEYWORDS:

• Opinion formation models
• French-DeGroot model
• bounded confidence models

Acknowledgments

The authors are grateful to anonymous reviewers for their invaluable comments.

Notes

¹ https://scikit-learn.org/stable/modules/generated/sklearn.decomposition.TruncatedSVD.html

² https://scikit-learn.org/stable/modules/generated/sklearn.linear_model.LogisticRegression.html

Additional information

Funding

This work was supported by the Russian Foundation for Basic Research, project no. 18-29-22042.