针对弱关系网络中由于强关系缺失而导致危险言论引导者检测指向结果弱的问题,提出了一种在弱关联社交网络中检测危险言论引导者的方法。该方法基于多表征融合,通过确定用户的言论情感和敏感性,评估用户的专业性表征,将用户言论作为节点构建社交环境网络,并利用图神经网络计算文本共现关系及用户在社交环境网络的重要性指数,融合用户在社交网络中的活跃度表征来检测社交网络中的危险言论领导者。结果表明所提出的方法相较于基线方法具有更好的性能。
Aiming at the problem of detecting dangerous speech leaders in weak relational networks, this paper proposes a method for detecting dangerous speech leaders in weakly connected social networks. The proposed method is based on multi-representation fusion. It determines the sentiment and sensitivity of users' speech to evaluate their professionalism representation. Users' speech is used as nodes to construct a social environment network, and a graph neural network is employed to calculate text co-occurrence relationships and the importance index of users within this network. The user activity representation in social network is then integrated to detect dangerous speech leaders. The results show that the proposed method has better performance compared to baselines.
[1] LAZARSFELD P F, KATZ E. Personal Influence: the Part Played by People in the Flow of Mass Communications[M]. Glencoe, ILL: The Free Press, 1955.
[2] 孙羽, 裘江南. 基于网络分析和文本挖掘的意见领袖影响力研究[J]. 数据分析与知识发现, 2022, 6(1): 69-79.
SUN Y, QIU J N. Studying opinion leaders with network analysis and text mining[J]. Data Analysis and Knowledge Discovery, 2022, 6(1): 69-79.
[3] BUERGER C, BENESCH S. Dangerous speech as an atrocity early warning indicator: measuring changing conflict dynamics[J]. Genocide Studies and Prevention: an International Journal, 2024, 18(1): 84-95.
[4] DASH S, GROVER R, SHEKHAWAT G, et al. Insights into incitement: a computational perspective on dangerous speech on twitter in India[C]. Proceedings of the 5th ACM SIGCAS/SIGCHI Conference on Computing and Sustainable Societies (COMPASS '22). New York, NY, USA: Association for Computing Machinery, 2022: 103-121.
[5] SONIA K, SHARMA M, BAJAJ M. A review on opinion leader detection and its applications[C]. 2022 7th International Conference on Communication and Electronics Systems (ICCES). Coimbatore, India: IEEE, 2022: 1645-1651.
[6] JAIN L, KATARYA R, SACHDEVA S. Opinion leader detection using whale optimization algorithm in online social network[J]. Expert Systems with Applications, 2020, 142: 113016.
[7] BARNIDGE M, XENOS M A. Social media news deserts: digital inequalities and incidental news exposure on social media platforms[J]. New Media & Society, 2024, 26(1): 368-388.
[8] TAHA K, SHOUFAN A, TAHA A. Unlocking criminal hierarchies: a survey, experimental, and comparative exploration of techniques for identifying leaders within criminal networks[DB/OL].[2024-07-30]. https://arXiv:2402.03355,2024.
[9] VASWANI A, SHAZEER N, PARMAR N, et al. Attention is all you need[C]. Advances in Neural Information Processing Systems. Long Beach, USA: Curran Associates, Inc, 2017: 5998-6008.
[10] USAMA M, AHMAD B, SONG E, et al. Attention-based sentiment analysis using convolutional and recurrent neural network[J]. Future Generation Computer Systems, 2020, 113: 571-578.
[11] YAN H, DENG B, LI X, et al. TENER: adapting transformer encoder for named entity recognition[DB/OL].[2024-08-10].https://arXiv:1911.04474,2019.
[12] 肖国庆, 李雪琪, 陈玥丹,等. 大规模图神经网络研究综述[J]. 计算机学报, 2024, 47(1): 148-171.
XIAO G Q, LI X Q, CHEN Y D, et al. A survey of large-scale graph neural networks[J]. Chinese Journal of Computers, 2024, 47(1): 148-171.
[13] 杨洁祎, 董一鸿, 钱江波. 基于图神经网络的小样本学习方法研究进展[J]. 计算机研究与发展, 2024, 61(4): 856-876.
YANG J Y, DONG Y H, QIAN J B. Research progress of few-shot learning methods based on graph neural networks[J]. Journal of Computer Research and Development, 2024, 61(4): 856-876.
[14] ZHAO Y, KOU G. Opinion evolution of a social group with extreme opinion leaders[C]. International Conference on Computers Communications and Control. Oradea, Romania: IEEE, 2016: 70-74.
[15] RODRIGO S G, ROSÁRIO M O, LINA O. Theoretical derivation of interval principal component analysis[J]. Information Sciences, 2023, 621: 227-247.
[16] CAO H, WANG J, WANG Z. Opinion leaders discovery in social networking site based on the theory of propagation probability[C]. 2nd IEEE Advanced Information Management, Communicates, Electronic and Automation Control Conference. Xi’an, China: IEEE, 2018: 700-704.
[17] TAN M, SANTOS C, XIANG B, et al. Lstm-based deep learning models for non-factoid answer selection[DB/OL].[2024-08-09].https://arXiv:1511.04108,2015.
[18] 张家伟, 高冠东, 肖珂等. 分层注意网络和TextCNN联合建模的暴力犯罪分级算法[J]. 计算机应用, 2024, 44(2): 403-410.
ZHANG J W, GAO G D, XIAO K, et al. Violent crime hierarchy algorithm by joint modeling of improved hierarchical attention network and TextCNN[J]. Journal of Computer Applications, 2024, 44(2): 403-410.
[19] YANG X, LI Y, LI Q, et al. Temporal-spatial three-way granular computing for dynamic text sentiment classification[J]. Information Sciences, 2022, 596: 551-566.
[20] KERAMATFAR A, AMIRKHANI H, BIDGOLY A J. Modeling tweet dependencies with graph convolutional networks for sentiment analysis[J]. Cognitive Computation, 2022, 14(6): 2234-2245.
[21] XIE Y, HUANG X, XIE X, et al. A fake news detection framework using social user graph[C]. Proceedings of the 2020 2nd International Conference on Big Data Engineering. Amsterdam, Netherlands: ACM, 2020: 55-61.
[22] COHEN I, HUANG Y, CHEN J, et al. Pearson Correlation Coefficient[M]. Berlin, Heidelberg: Springer, Noise reduction in Speech Processing, 2009.
