基于传播时空特性的社交网络灰帽用户检测

doi:10.19678/j.issn.1000-3428.0059636

摘要/Abstract

摘要： 社交网络灰帽用户虽然种类多样又善于伪装，但因最终目的都是扩大自身影响力，故在交互行为上具有共同特性，即在其用户生成内容(user generated content, UGC)或参与他人 UGC 传播过程中与白帽用户相比表现出明显差异。基于此，为了解决社交网络灰帽用户因极易隐藏且类型多样导致的检测算法适用性低的问题，本文提出一种基于传播时空特性的社交网络检测机制—DSTC(Diffusion Spatiotemporal Characteristics)。首先，挖掘社交网络白帽/灰帽用户各自的 UGC 在传播时序上的不同特性；而后，通过构建 UGC 传播网络来度量白帽/灰帽用户在传播空间上的不同特性；基于挖掘出的传播时空特性，设计了社交网络用户检测机制 DSTC，融合时空传播特性并调节权重比例提高分类性能。实验结果表明，与几类主流灰帽用户检测算法相比，DSTC 在多个数据集上的 Accuracy 值最高提升 26.08%，AUC 值有 0.008%~30.54%的大幅提升，表明 DSTC 能有效检测不同类型潜藏灰帽用户。

Abstract: Although social network gray hat users are diverse and good at disguising, their ultimate goal is to expand their own influence, so they have common characteristics in their interactive behavior. That is, in their user generated content (UGC) or participate in the spread of UGC by others, there are obvious differences in the diffusion process compared with white hat users. Based on this, in order to solve the problem of low applicability of detection algorithms due to the easy concealment of social network gray hat users and their diverse types, this paper proposes a social network detection mechanism based on the spreading spatiotemporal characteristics of propagation—DSTC (Diffusion Spatiotemporal Characteristics). Firstly, exploring the different characteristics of the dissemination timing of the respective UGC of the white/gray hat users of social networks; then, by constructing a UGC dissemination network to measure the different characteristics of the white/gray hat users in the dissemination space; based on the dissemination discovered spatiotemporal characteristics, a social network user detection mechanism DSTC is designed, which integrates spatiotemporal propagation characteristics and adjusts the weight ratio to improve classification performance. Experimental results show that, compared with several mainstream gray hat user detection algorithms, the Accuracy value of DSTC on multiple data sets is increased by 26.08%, and the AUC value is greatly increased from 0.008% to 30.54%, indicating that DSTC can effectively detect different types of potential gray hat users.

何欢 , 朱焱 , 李春平. 基于传播时空特性的社交网络灰帽用户检测[J]. 计算机工程, doi: 10.19678/j.issn.1000-3428.0059636.

HE Huan , ZHU Yan , LI Chunping . Grey Hat User Detection in Social Network Based on Spatiotemporal Characteristics of Diffusion[J]. Computer Engineering, doi: 10.19678/j.issn.1000-3428.0059636.

参考文献

[1] W. YUE, C. Xiaoliang, J. Zhongyuan. Survey on Predicting Popularity of Information in Microblogs. 吴越, 陈晓亮, 蒋忠远. 微博信息流行度预测研究综述[J]. 西华大学学报（自然科学版）, 2017(1).
[2] B. Erçahin, Ö. Aktaş, D. Kilinç, et al. Twitter fake account detection[C]// Proc of the 2nd International Conference on Computer Science and Engineering. Piscataway, NJ: IEEE Press, 2017: 388–392.
[3] RAYMOND Y K, STEPHEN L, LIAO S Y. Text mining and probabilistic language modeling for online review spam detection[J]. ACM Transactions on Management Information Systems, Volume 2, 2011, 2(4): 25:1-25:30.
[4] ZHANG X, ZHU S, LIANG W. Detecting spam and promoting campaigns in the Twitter social network[C]//Proc of the 12th IEEE International Conference on Data Mining. Brussels, ICDM 2012: 1194-1199.
[5] Neil Z Gong, Mario F, Prateek M. Sybil Belief: A Semi-Supervised Learning Approach for Structure-Based Sybil Detection[J]. IEEE Transactions on Information Forensics and Security, Volume 9(6), 2014: 976-987.
[6] N. Tran, J. Li, L. Subramanian, and S. S. Chow. Optimal sybil-resilient node admission control[J]. INFOCOM, 2011 Proceedings IEEE, 2011, pp. 3218–3226.
[7] J. Jia, B. Wang, et al. Random walk based fake account detection in online social networks[C]// Proc of the Dependable Systems and Networks. DSN 2017: 273–284.
[8] B. Wang, L. Zhang, and N. Z. Gong. Sybil scar: Sybil detection in online social networks via local rule based propagation[J]. INFOCOM 2017-IEEE Conference on Computer Communications, 2017, pp. 1–9.
[9] Z. Yang, C. Wilson, X. Wang, et al. Uncovering social network sybils in the wild[J]// Proc of the ACM Transactions on Knowledge Discovery from Data (TKDD), Volume 9(1), 2014: 2:1-2:29.
[10] V. Sridharan, V. Shankar, and M. Gupta. Twitter games: How successful spammers pick targets[C]// Proc of the 28th Annual Computer Security Applications Conference. ACM, 2012: 389–398.
[11] Y. Boshmaf, I. Muslukhov, K. Beznosov, et al. The social bot network: when bots socialize for fame and money[C]// Proc of the 27th annual computer security applications conference. ACM, 2011, pp. 93–102.
[12] D. Koll, M. Schwarzmaier , et al. Thank you for being a friend: An attacker view on online-social-network-based sybil defenses[C]// Proc of the 37th IEEE International Conference on Distributed Computing Systems Workshops. ICDCS Workshops, 2017: 157-162
[13] S. Effendy and R. H. Yap. The strong link graph for enhancing sybil defenses[C]// Prof the 37th IEEE International Conference on Distributed Computing Systems. ICDCS 2017: 944-954.
[14] Zhang X, Xie H, Lui J C S. Sybil Detection in Social-Activity Networks: Modeling, Algorithms and Evaluations[C]// Prof of the 26th IEEE International Conference on Network Protocols. ICNP 2018: 44-54.
[15] Stefano C, Roberto D, et al. Social Fingerprinting: Detection of Spambot Groups Through DNA-Inspired Behavioral Modeling[J].IEEE Transactions on Dependable and Secure Computing, Volume 15(4), 2018: 561-576.
[16] M. Mateen, M. A. Iqbal, M. Aleem, and M. A. Islam. A hybrid approach for spam detection for Twitter[C]// Proc of the 14th Int. Bhurban Conf. Appl. Sci. Technol. (IBCAST), Jan. 2017, pp. 466–471.
[17] Chaozhuo L, Senzhang W, Lifang H, et.al. SSDMV: Semi-supervised Deep Social Spammer Detection by Multi-View Data Fusion[C]// Prof of the 18th IEEE International Conference on Data Mining. Singapore: ICDM, 2018: 247-256.
[18] Liu Y , Wu B, et al. SDHM: A hybrid model for spammer detection in Weibo[C]// Proc of the 2014 IEEE/ACM International Conference on Advances in Social Networks Analysis and Mining (ASONAM). ACM, 2014.
[19] Grover A, Leskovec J. node2vec: Scalable Feature Learning for Networks[C]// Proc of the 2016 Acm Sigkdd International Conference on Knowledge Discovery & Data Mining. 2016.
[20] Lee, Kyumin, Brian David Eoff, James Caverlee. Seven Months with the Devils: A Long-Term Study of Content Polluters on Twitter[C]// Prof of the 5th International AAAI Conference on Weblogs and Social Media (ICWSM), Barcelona, July 2011.
[21] Yang, Kai-Cheng, Onur Varol, et.al. Scalable and Generalizable Social Bot Detection through Data Selection[C]// Prof of the Thirty-Fourth AAAI Conference on Artificial Intelligence. AAAI press: 2020: 1096-1103.
[22] Cresci, S., Di Pietro, R., Petrocchi, et.al. Fame for sale: efficient detection of fake Twitter followers[J]. Decision Support Systems, Volume 80,2015: 56-71.