Details

Autor(en) / Beteiligte

• Zhang, Xian-Jie
• Wang, Jing
• Ma, Xiao-Jing
• Ma, Chuang
• Kan, Jia-Qian
• Zhang, Hai-Feng

Titel

Influence maximization in social networks with privacy protection

Ist Teil von

• Physica A, 2022-12, Vol.607, p.128179, Article 128179

Ort / Verlag

Elsevier B.V

Erscheinungsjahr

2022

Quelle

Alma/SFX Local Collection

Beschreibungen/Notizen

• With the explosive development of online social network platforms, how to find a small subset of users (seed nodes) across multiple social networks to maximize the spread of information is of great significance. In reality, different platforms need to consider not only the commercial value of data, but also the protection of data privacy. In this situation, these multiple social platforms can be treated as a system of “multiple private social networks”, which naturally arises a new problem: how to maximize the spread of information in multiple private social networks without breaking the protocol of privacy protection. In view of this, we propose an HE-IM algorithm to solve the problem from the perspective of cryptography. Specifically, we use the homomorphic encryption security protocol and the third-party servers to encrypt and decrypt the influence of nodes in each private network and update the set of seed nodes. The experimental results demonstrate that, by cooperatively fusing information from different private networks in a secret manner, our method can effectively find influential seed nodes to maximize influence in multiple private social networks. The performance of our method in maximizing influence range is much better than that of the baseline methods only considering the structure of single private network. Therefore, the method provides a new way for collaborative search of influential seed nodes in multiple private social networks. •An all-new problem is defined: how to maximize the spread of information in social networks with privacy protection.•An HE-IM method based on the homomorphic encryption protocol is designed to solve the problem.•The method can collaboratively fuse the structural information of different private networks in a secret manner.•Our method outperforms the baseline methods in maximizing the spread of influence.