Cover not available
Article published In:
The Discourse of Terrorism
Edited by Encarnación Hidalgo-Tenorio and Juan L. Castro
[Pragmatics and Society 13:3] 2022
► pp. 532–554

A semi-supervised algorithm for detecting extremism propaganda diffusion on social media

| University of Granada
| University of Zaragoza
| University of Granada
| University of Granada
Extremist online networks reportedly tend to use Twitter and other Social Networking Sites (SNS) in order to issue propaganda and recruitment statements. Traditional machine learning models may encounter problems when used in such a context, due to the peculiarities of microblogging sites and the manner in which these networks interact (both between themselves and with other networks). Moreover, state-of-the-art approaches have focused on non-transparent techniques that cannot be audited; so, despite the fact that they are top performing techniques, it is impossible to check if the models are actually fair. In this paper, we present a semi-supervised methodology that uses our Discriminatory Expressions algorithm for feature selection to detect expressions that are biased towards extremist content (Francisco and Castro 2020). With the help of human experts, the relevant expressions are filtered and used to retrieve further extremist content in order to iteratively provide a set of relevant and accurate expressions. These discriminatory expressions have been proved to produce less complex models that are easier to comprehend, and thus improve model transparency. In the following, we present close to 70 expressions that were discovered by using this method alongside the validation test of the algorithm in several different contexts.
Keywords: feature selection, discriminatory expressions, extremism, propaganda, social media, interpretability, text mining, microblogging
Article outline
  • Introduction
  • Theoretical background
    • What is a model and how do we train it?
    • How can we check that the model learned correctly?
    • Can models be interpreted by humans?
    • How do models deal with Natural Language documents?
    • Is it possible to reduce the dimensionality of the vector representation?
    • What are the reference filtering methods?
      • CHI2 (chi-square)
      • Information gain (IG)
      • Mutual information (MI)
      • Odds ratio (OR)
      • Expected cross entropy (ECE)
      • ANOVA F-value
      • Galavotti-sebastiani-simi coefficient (GSS)
    • Are filters going to help us comprehend models?
    • How can we be sure that this is the way to go?
  • Discriminatory expressions (DE)
    • Definition (Expression)
      • Definition (Discriminatory Expression)
  • Methodology
  • Experiments
  • Performance and comprehensibility tests
  • Application-related Tests
  • Results and discussion
  • Application-specific results
  • Limitations
  • Conclusions
  • Future work
  • Notes
  • References
Published online: 21 July 2022
DOI logo
https://doi.org/10.1075/ps.21009.fra
References (37)
References
Alharbi, Ahmed S. M., and Elise de Doncker. 2019. ‘Twitter Sentiment Analysis with a Deep Neural Network: An Enhanced Approach Using User Behavioral Information’. Cognitive Systems Research 541: 50–61. DOI logoGoogle Scholar
Al-Salemi, Bassam, Shahrul Azman Mohd Noah, and Mohd Juzaiddin Ab Aziz. 2016. ‘RFBoost: An Improved Multi-Label Boosting Algorithm and Its Application to Text Categorisation’. Knowledge-Based Systems 1031 (July): 104–17. DOI logoGoogle Scholar
Alvari, Hamidreza, Soumajyoti Sarkar, and Paulo Shakarian. 2019. ‘Detection of Violent Extremists in Social Media’. ArXiv:1902.01577 [Cs] , February. [URL]. DOI logo
Ashktorab, Zahra, Christopher Brown, Manojit Nandi, and Aron Culotta. 2014. ‘Tweedr: Mining Twitter to Inform Disaster Response.’ In ISCRAM.Google Scholar
Benigni, Matthew C., Kenneth Joseph, and Kathleen M. Carley. 2017. ‘Online Extremism and the Communities That Sustain It: Detecting the ISIS Supporting Community on Twitter’. PLOS ONE 12 (12): e0181405. DOI logoGoogle Scholar
Caropreso, Maria Fernanda, Stan Matwin, and Fabrizio Sebastiani. 2001. ‘A Learner-Independent Evaluation of the Usefulness of Statistical Phrases for Automated Text Categorization’, 151.Google Scholar
Cowan, Nelson. 2001. ‘The Magical Number 4 in Short-Term Memory: A Reconsideration of Mental Storage Capacity’. The Behavioral and Brain Sciences 24 (1): 87–114; discussion 114–185. DOI logoGoogle Scholar
Deng, Xuelian, Yuqing Li, Jian Weng, and Jilian Zhang. 2019. ‘Feature Selection for Text Classification: A Review’. Multimedia Tools and Applications 78 (3): 3797–3816. DOI logoGoogle Scholar
Ding, Jianli, and Liyang Fu. 2018. ‘A Hybrid Feature Selection Algorithm Based on Information Gain and Sequential Forward Floating Search’. Journal of Intelligent Computing 9 (3): 93. DOI logoGoogle Scholar
FAT/ML. n.d. ‘Principles for Accountable Algorithms and a Social Impact Statement for Algorithms’. Accessed 8 January 2019. [URL]
Forman, George. 2003. ‘An Extensive Empirical Study of Feature Selection Metrics for Text Classification [J]’. Journal of Machine Learning Research – JMLR 31 (March).Google Scholar
Francisco, Manuel, and Juan Luis Castro. 2020. ‘Discriminatory Expressions to Produce Interpretable Models in Microblogging Context’. ArXiv:2012.02104 [Cs] , November. [URL]
Galavotti, Luigi, Fabrizio Sebastiani, and Maria Simi. 2000. ‘Experiments on the Use of Feature Selection and Negative Evidence in Automated Text Categorization’. In Research and Advanced Technology for Digital Libraries, edited by José Borbinha and Thomas Baker, 59–68. Lecture Notes in Computer Science. Berlin, Heidelberg: Springer. DOI logoGoogle Scholar
Go, Alec, Richa Bhayani, and Lei Huang. 2009. ‘Twitter Sentiment Classification Using Distant Supervision’. Processing 1501 (January).Google Scholar
Harris, Zellig S. 1954. ‘Distributional Structure’. Word 10 (2–3): 146–62. DOI logoGoogle Scholar
Kotzias, Dimitrios, Misha Denil, Nando de Freitas, and Padhraic Smyth. 2015. ‘From Group to Individual Labels Using Deep Features’. In KDD ’15. DOI logoGoogle Scholar
Kubat, Miroslav. 2017. An Introduction to Machine Learning. Cham: Springer International Publishing. DOI logoGoogle Scholar
Largeron, Christine, Christophe Moulin, and Mathias Géry. 2011. ‘Entropy Based Feature Selection for Text Categorization’. In ACM Symposium on Applied Computing, edited by William C. Chu, W. Eric Wong, Mathew J. Palakal, and Chih-Cheng Hung, 924–28. TaiChung, Taiwan: ACM. DOI logoGoogle Scholar
Miller, George A. 1956. ‘The Magical Number Seven, plus or Minus Two: Some Limits on Our Capacity for Processing Information’. Psychological Review 63 (2): 81–97. DOI logoGoogle Scholar
Misangyi, Vilmos F., Jeffery A. LePine, James Algina, and Jr Francis Goeddeke. 2016. ‘The Adequacy of Repeated-Measures Regression for Multilevel Research: Comparisons With Repeated-Measures ANOVA, Multivariate Repeated-Measures ANOVA, and Multilevel Modeling Across Various Multilevel Research Designs’. Organizational Research Methods, June. DOI logoGoogle Scholar
O’Dair, M., and A. Fry. 2019. ‘Beyond the Black Box in Music Streaming: The Impact of Recommendation Systems upon Artists’. Popular Communication. DOI logoGoogle Scholar
Periñán-Pascual, Carlos, and Francisco Arcas-Túnez. 2019. ‘Detecting Environmentally-Related Problems on Twitter’. Biosystems Engineering, Intelligent Systems for Environmental Applications, 1771 (January): 31–48. DOI logoGoogle Scholar
Phillips, Avery. 2018. ‘The Moral Dilemma of Algorithmic Censorship’. Becoming Human: Artificial Intelligence Magazine. 27 August 2018. [URL]
Rudin, Cynthia. 2018. ‘Please Stop Explaining Black Box Models for High Stakes Decisions’. ArXiv:1811.10154 [Cs, Stat] , November. [URL]
Rutkowski, Leszek, Ryszard Tadeusiewicz, Lofti A. Zadeh, and Jacek M. Zurada. 2008. Artificial Intelligence and Soft Computing – ICAISC 2008: 9th International Conference Zakopane, Poland, June 22–26, 2008, Proceedings. Springer Science & Business Media. DOI logoGoogle Scholar
Senthil, Kumar B. and Varma E. Bhavitha. 2016. ‘A Different Type of Feature Selection Methods for Text Categorization on Imbalanced Data’ 5 (9): 7.Google Scholar
Sparck-Jones, Karen. 1972. ‘A Statistical Interpretation of Term Specificity and Its Application in Retrieval’. Journal of Documentation 28 (1): 11–21. DOI logoGoogle Scholar
Twitter Inc. 2019. ‘Q1 2019 Earning Report’. [URL]
‘Twitter Usage Statistics – Internet Live Stats’. 2013. 2013. [URL]
Villena-Román, Julio, Sara Lana-Serrano, Eugenio Martínez-Cámara, and José Carlos González-Cristóbal. 2013. ‘TASS – Workshop on Sentiment Analysis at SEPLN’. Procesamiento del Lenguaje Natural 50 (0): 37–44.Google Scholar
Wang, Hao, Dogan Can, Abe Kazemzadeh, François Bar, and Shrikanth Narayanan. 2012. ‘A System for Real-Time Twitter Sentiment Analysis of 2012 U.S. Presidential Election Cycle’. In Proceedings of the ACL 2012 System Demonstrations, 115–20. ACL ’12. Stroudsburg, Penn.: Association for Computational Linguistics. [URL]
Wu, Guohua, Liuyang Wang, Nailiang Zhao, and Hairong Lin. 2015. ‘Improved Expected Cross Entropy Method for Text Feature Selection’. In 2015 International Conference on Computer Science and Mechanical Automation (CSMA), 49–54. DOI logoGoogle Scholar
Xu, Yan, Gareth Jones, Jintao Li, Bin Wang, and Chunming Sun. 2007. ‘A Study on Mutual Information-Based Feature Selection for Text Categorization’. Journal of Computational Information Systems 31 (March).Google Scholar
Xue, Bing, Mengjie Zhang, and Will Browne. 2013. ‘Particle Swarm Optimization for Feature Selection in Classification: A Multi-Objective Approach’. IEEE Transactions on Cybernetics 431 (December): 1656–71. DOI logoGoogle Scholar
Zhao, Z., M. Gao, J. Yu, Y. Song, X. Wang, and M. Zhang. 2018. ‘Impact of the Important Users on Social Recommendation System’. Lecture Notes of the Institute for Computer Sciences, Social-Informatics and Telecommunications Engineering, LNICST 2521: 425–34. DOI logoGoogle Scholar
Zheng, Hai-Tao, Zhe Wang, Wei Wang, Arun Kumar Sangaiah, Xi Xiao, and Congzhi Zhao. 2018. ‘Learning-Based Topic Detection Using Multiple Features’. Concurrency and Computation-Practice & Experience 30 (15): e4444. DOI logoGoogle Scholar
Zheng, Zhaohui, Xiaoyun Wu, and Rohini Srihari. 2004. ‘Feature Selection for Text Categorization on Imbalanced Data’. ACM SIGKDD Explorations Newsletter 6 (1): 80–89. DOI logoGoogle Scholar
Cited by (1)

Cited by one other publication

Wang, Mengdi, Xiaobing Peng & Liang Zhuang
2023. Publicity governance in contingency management during the COVID-19 pandemic in China: A “Government-Society” perspective. PLOS ONE 18:11 ► pp. e0293210 ff. DOI logo

This list is based on CrossRef data as of 5 july 2024. Please note that it may not be complete. Sources presented here have been supplied by the respective publishers. Any errors therein should be reported to them.