LDPCD: A Novel Method for Locally Differentially Private Community Detection.

Zhejian Zhang,Bai Yuan Ding

doi:10.1155/2022/4080047

Zhejian Zhang, Bai Yuan Ding

Open Access

https://doi.org/10.1155/2022/4080047

Copy DOI

Abstract

As one of the cores of data analysis in large social networks, community detection has become a hot research topic in recent years. However, user's real social relationship may be at risk of privacy leakage and threatened by inference attacks because of the semitrusted server. As a result, community detection in social graphs under local differential privacy has gradually aroused the interest of industry and academia. On the one hand, the distortion of user's real data caused by existing privacy-preserving mechanisms can have a serious impact on the mining process of densely connected local graph structure, resulting in low utility of the final community division. On the other hand, private community detection requires to use the results of multiple user-server interactions to adjust user's partition, which inevitably leads to excessive allocation of privacy budget and large error of perturbed data. For these reasons, a new community detection method based on the local differential privacy model (named LDPCD) is proposed in this paper. Due to the introduction of truncated Laplace mechanism, the accuracy of user perturbation data is improved. In addition, the community divisive algorithm based on extremal optimization (EO) is also reﬁned to reduce the number of interactions between users and the server. Thus, the total privacy overhead is reduced and strong privacy protection is guaranteed. Finally, LDPCD is applied in two commonly used real-world datasets, and its advantage is experimentally validated compared with two state-of-the-art methods.

Highlights

Due to the rapid development of Internet technology, APPs with various functions have brought great convenience to the daily interaction among users
Some general but necessary data are enumerated, which give a brief outline of the comparison between the classical extremal optimization (EO) algorithm and LDPCD when they are separately employed in the datasets. en, deeper discussions about the influence of parameters and the contrast result of LDPCD with two state-of-the-art methods will be elaborated according to the figures
In the framework of LDPCD, the truncated Laplace mechanism with local differential privacy is employed to enhance the accuracy of user perturbation data

Summary

Introduction

Due to the rapid development of Internet technology, APPs with various functions have brought great convenience to the daily interaction among users After integration, these relationship data of users can be exploited to build social graphs, from which service providers can mine valuable information such as frequent subgraphs [1, 2], average path length among users [3], and the community structure of users [4,5,6]. As an important manifestation of topological features, community structure has a significant guiding role in creating synthetic social graphs [9]. Is privacy scandal indicates that one of the main problems to be solved in the community mining of social networks is the privacy protection during the collection of user relationship data In 2018, Facebook was accused of leaking tens of millions of user personal information to the UK-based thirdparty firm Cambridge Analytica [10]. is privacy scandal indicates that one of the main problems to be solved in the community mining of social networks is the privacy protection during the collection of user relationship data

Objectives

Methods

Results

Conclusion