Efficient Intervention in the Spread of Misinformation in Social Networks

• Published in: IEEE access Vol. 12; pp. 133489 - 133498
• Main Authors: Sakiyama, Takumi, Nakajima, Kazuki, Aida, Masaki
• Format: Journal Article
• Language: English
• Published: Piscataway IEEE 2024; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Algorithms; Approximation algorithms; Computation; Diffusion models; Diffusion rate; Effectiveness; Fake news; False information; Greedy algorithms; Heuristic algorithms; Heuristic methods; influence maximization; Integrated circuit modeling; Intervention; misinformation; online social networks; Social networking (online); Social networks; Topology
• ISSN: 2169-3536; 2169-3536
• DOI: 10.1109/ACCESS.2024.3459830

The spread of misinformation through social media has often impacted public opinion and precipitated social disruption. In response to such real-world instances, strategies to intervene in the diffusion of misinformation among individuals have been developed. A basic intervention approach involves disseminating correct information to the largest possible population in a social network. In the present study, we investigate the problem of identifying a set of seed individuals that maximize the spread of correct information throughout a social network where the diffusion of misinformation has occurred. We find that a greedy algorithm for this problem lacks theoretical guarantees regarding the approximation ratio of the spread of correct information and requires considerable computation time in large-scale networks. To address this efficiency issue, we propose a heuristic algorithm called MCEaSyIM. The proposed algorithm exploits stochastic properties of the simultaneous diffusion of misinformation and correct information on the network to efficiently compute the seed individuals that approximately maximize the expected diffusion of correct information. Our algorithm achieves a comparable spread of correct information in a considerably faster computation time compared with the greedy algorithm. In addition, our algorithm achieves a broader spread of correct information in a reasonable computation time than topology-based heuristics, such as the method that selects seed individuals with the largest out-neighbors that have not been activated with misinformation. Moreover, we found that the effectiveness of our algorithm in spreading correct information is partially associated with the degree assortativity of the network. We expect that our algorithm will serve as an efficient and effective intervention strategy in combating the spread of misinformation in social networks.