Social media platforms have emerged as a hub for political and social interactions, and analyzing the polarization of opinions has been gaining attention. In this study, we have proposed a measure to quantify polarization on social networks. The proposed metric, unlike state-of-the-art methods, does not assume a two-opinion case and applies to multiple opinions. We tested our metric on different networks with a multi-opinion scenario and varying degrees of polarization. The scores obtained from the proposed metric were comparable to state-of-the-art methods on binary opinion-based benchmark networks. The technique also differentiated among networks with different levels of polarization in a multi-opinion scenario. We also quantified polarization in a retweet network obtained from Twitter regarding the usage of drugs like hydroxychloroquine or chloroquine in treating COVID-19. Our metric indicated a high level of polarized opinions among the users. These findings suggest uncertainty among users in the benefits of using hydroxychloroquine and chloroquine drugs to treat COVID-19 patients.
@article{arxiv.2204.08697,
title = {A Multi-Opinion Based Method for Quantifying Polarization on Social Networks},
author = {Maneet Singh and S. R. S. Iyengar and Rishemjit Kaur},
journal= {arXiv preprint arXiv:2204.08697},
year = {2022}
}