MU Yunxiang,SHENG Zhiwei,LU Jiazhong.Research on the Algorithm of Online Water Army Recognition based on Minimax Game[J].Journal of Chengdu University of Information Technology,2023,38(03):306-313.[doi:10.16836/j.cnki.jcuit.2023.03.009]
基于极小极大博弈的水军识别算法研究
- Title:
- Research on the Algorithm of Online Water Army Recognition based on Minimax Game
- 文章编号:
- 2096-1618(2023)03-0306-08
- 分类号:
- TP393
- 文献标志码:
- A
- 摘要:
- 随着互联网的发展,用户越来越多地在线上完成购物、订餐,并倾向于先参考线上评论。评论对用户决策的重要导向作用催生了网络水军。网络水军会为了自身利益或其他不良动机,发布与实际体验不相符的评价,且会随时调整自己的策略来逃避平台的识别。现提出一个基于行为特征的水军识别算法(FBS),并将FBS加入到极小极大博弈,在这个博弈中,水军与识别器相互竞争,将博弈转换为两个相互依赖的马尔可夫决策过程,不断优化各自的策略,最终得到一个当前场景下最优的识别器。与当前先进的水军识别算法对比,性能有了明显提升,在公开数据集YelpChi上实际效应可以达到3.69。
- Abstract:
- With the Internet’s development, more and more users complete shopping and dining online. At the same time, the public will also tend to refer to online comments first. The important guiding role of comments in user decision-making gave birth to the network Navy. For its interests or other bad motives, the online Navy will release evaluations that are inconsistent with the experience. And the Navy will adjust its strategy at any time to to avoid the platform’s recognition. This paper proposes a behavior based Navy recognition algorithm(FBS), and adds FBS to the minimax game. In this game, the Navy and the recognizer compete, convert the game into two interdependent Markov decision-making processes, constantly optimize their strategies, and finally get an optimal recognizer in the current scene. Compared with the current advanced navy recognition algorithm, the neutral energy has been significantly improved, the actual effect based on the public dataset yelpchi can reach 3.69.
参考文献/References:
[1] Mukherjee A,Venkataraman V,Liu B,et al.What yelp fake review filter might be doing?[C].Proceedings of the International AAAI Conference on Web and Social Media.2013,7(1):409-418.
[2] Rayana S,Akoglu L.Collective opinion spam detection:Bridging review networks and metadata[C]. Proceedings of the 21th acm sigkdd international conference on knowledge discovery and data mining. 2015: 985-994.
[3] Luca M. Reviews, reputation, and revenue: The case of Yelp. com[J]. Com(March 15, 2016). Harvard Business School NOM Unit Working Paper. 2016, 12(16): 175-216.
[4] Leone, Michele, Sum ed ha, et al. Clustering by soft-constraint affinity propagation: applications to gene-expression data.[J]. Bioinformatics, 2007, 23(20): 2708-2715.
[5] Kotsiantis S B. Supervised Machine Learning: A Review of Classification Techniques[J]. Informatica, 2007, 31: 249-268.
[6] Dou Y, Ma G, Yu P S, et al. Robust spammer detection by nash reinforcement learning[C]. Proceedings of the 26th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining. 2020: 924-933.
[7] Mccord M, Chuah M. Spam Detection on Twitter Using Traditional Classifiers[C]. Autonomic & Trusted Computing-international Conference. DBLP, 2011: 175-186.
[8] Li J, Ott M, Cardie C, et al. Towards a general rule for identifying deceptive opinion spam[C]. Proceedings of the 52nd Annual Meeting of the Association for Computational Linguistics(Volume 1: Long Papers). 2014: 1566-1576.
[9] Noekhah S, binti Salim N, Zakaria N H. Opinion spam detection: Using multi-iterative graph-based model[J]. Information Processing & Management, 2020, 57(1): 102140.
[10] Wang X,Kang L,Zhao J. Handling Cold-Start Problem in Review Spam Detection by Jointly Embedding Texts and Behaviors[C]. Meeting of the Association for Computational Linguistics. 2017: 366-376.
[11] Hooi B, Song H A, Beutel A, et al. Fraudar: Bounding graph fraud in the face of camouflage[C]. Proceedings of the 22nd ACM SIGKDD international conference on knowledge discovery and data mining. 2016: 895-904.
[12] Wang B, Gong N Z, Fu H. GANG: Detecting fraudulent users in online social networks via guilt-by-association on directed graphs[C]. 2017 IEEE International Conference on Data Mining(ICDM). IEEE, 2017: 465-474.
[13] Si-Cai H U, Sun J P, Sheng-Gen J U, et al. Chinese emotion feature selection method based on the extended emotion dictionary and the chi-square model[J]. Journal of Sichuan University(Natural Science Edition), 2019, 56(1): 37-44.
[14] Shah N, Beutel A, Gallagher B, et al. Spotting suspicious link behavior with fbox: An adversarial perspective[C]. 2014 IEEE International conference on data mining. IEEE, 2014: 959-964.
[15] 孙文. 网络新闻评论用户行为分析及水军识别方法研究[D].杭州:杭州电子科技大学,2019.
[16] 任亚峰, 姬东鸿, 张红斌, 等. 基于 PU 学习算法的水军的虚假评论识别研究[J]. 计算机研究与发展, 2015, 52(3): 639-648.
[17] Gatterbauer W, S Günnemann,Koutra D, et al. Linearized and Single-Pass Belief Propagation[J]. Proceedings of the Vldb Endowment, 2014, 8(5): 581-592.
[18] Kaghazgaran P, Caverlee J, Squicciarini A. Combating crowdsourced review manipulators: A neighborhood-based approach[C]. Proceedings of the Eleventh ACM International Conference on Web Search and Data Mining. 2018: 306-314.
[19] Miller F P,A F Vandome,J Mcbrewster.Amazon Mechanical Turk.[C].Alphascript Publishing.2021:308-331.
[20] Mukherjee A, Kumar A, Liu B, et al. Spotting opinion spammers using behavioral footprints[C]. Proceedings of the 19th ACM SIGKDD international conference on Knowledge discovery and data mining. 2013: 632-640.
[21] Y Chen,C Lou.Research on the formation path of fake reviews of online goods[J], Modern Intelligence. 2015, 8(10): 49-53.
备注/Memo
收稿日期:2022-07-16
基金项目:四川省科技计划资助项目(2021YFG0332)