Identifying influential nodes in social networks has drawn significant attention in the field of network science. However, most of the existing works request to know the complete structural information about networks, indeed, this information is usually sensitive, private and hard to obtain. Therefore, how to identify the influential nodes in networks without disclosing privacy is especially important. In this paper, we propose a privacy-preserving (named as HE-ranking) framework to identify influential nodes in networks based on homomorphic encryption (HE) protocol. The HE-ranking method collaboratively computes the nodes’ importance and protects the sensitive information of each private network by using the HE protocol. Extensive experimental results indicate that the method can effectively identify the influential nodes in the original networks than the baseline methods which only use each private network to identify influential nodes. More importantly, the HE-ranking method can protect the privacy of each private network in different parts.

识别社交网络中有影响力的节点引起了网络科学领域的广泛关注。然而，现有的大多数工作都要求了解网络的完整结构信息，事实上，这些信息通常是敏感的、私密的且难以获取。因此，如何在不泄露隐私的情况下识别网络中具有影响力的节点就显得尤为重要。在本文中，我们提出了一种隐私保护（称为 HE 排名）框架，用于基于同态加密（HE）协议来识别网络中有影响力的节点。HE-ranking方法通过使用HE协议协同计算节点的重要性并保护每个私有网络的敏感信息。大量的实验结果表明，与仅使用每个私有网络来识别影响节点的基线方法相比，该方法能够有效地识别原始网络中的影响节点。更重要的是，HE-ranking方法可以保护不同部分的每个私有网络的隐私。