Analysis of information-psychological impact detection methods in social networks

Authors

  • Valeriia Pokrovska Institute of special communication and information protection of National technical university of Ukraine “Igor Sikorsky Kyiv polytechnic institute”, Kyiv,, Ukraine https://orcid.org/0000-0002-1318-5521

DOI:

https://doi.org/10.20535/2411-1031.2020.8.1.218002

Keywords:

information-psychological impact, machine learning, neural networks, lexeme-oriented approach, sentiment analysis

Abstract

Methods with the help of which it is possible to carry out automatic content analysis in social networks for the detection information-psychological impact are analyzed. Based on the performed research, the features of the virtual communities functioning in social networks were determined. Virtual communities have become objects and tools of external information management and the information confrontation arena at different levels. They have become an ideal tool for information-psychological impact on the national interests of the state, society in the information, and cyberspace, in general. To prevent and counteract shocks in society, it is necessary to constantly monitor the presence of negative informational-psychological impact in communities to be able to resist it. Methods for detecting information-psychological impact include methods based on the use of lexemes and machine learning with a teacher, namely: support vector machine, the naive Bayes classifier, decision trees, the method of maximum entropy, and neural networks. Each of the analyzed methods has its own advantages and disadvantages, features of use, which must be taken into account when choosing a method for detecting information-psychological impact in social networks. Among the methods considered for automatic content analysis, the most effective method is a machine learning based on the use of neural networks. This method does not involve pre-processing of text, there is no need to create dictionaries, can classify into several categories. This allows identifying different types of information-psychological impact by training the network with new information. So, updates of content in social networks are taken into account. It has been established that unlike neural networks, the decision tree for detecting information-psychological impact cannot be used in practice. This limitation is due to the difficulty of maintaining incremental training. You can take a large amount of data and build a decision tree for it.  However, it is impossible to take into account new messages, because you will have to teach it from scratch every time.

Author Biography

Valeriia Pokrovska, Institute of special communication and information protection of National technical university of Ukraine “Igor Sikorsky Kyiv polytechnic institute”, Kyiv,

engineer at the cybersecurity
and application of information
systems and technologies
academic department

References

K. Molodetska-Grinchuk, “ Method for identifying signs of information impacts in social Internet services based on their content ”, Radio electronics, informatics, management, no. 2, pp. 117-126, 2017, doi: https://doi.org/10.15588/1607-3274-2017-2-13.

R. Grischuk, and K. Molodetska-Grinchuk, “ Methodology of building a system to ensure information security of the state in social Internet services ”, Information Protection, vol. 19, no. 4, pp. 25-262, 2017, doi: https://doi.org/10.18372/2410-7840.19.12204.

Most famous social network sites worldwide as of October 2018, ranked by number of active users (in millions). [Online]. Available: https://www.statista.com/statistics/272014/global-social-networks-ranked-by-number-of-users/. Accessed on: October 17, 2019.

А. Solovey. “Virtual communities as a special form of social integration”, Psychology, sociology and education, no. 6(33), pp. 33-41, 2014. [Online]. Available: http://psychology.snauka.ru/2014/06/3341. Accessed on: February 10, 2020.

V. Gorbulin, A. Dodonov, and D. Lande, Information operations and public safety: threats, counteraction, modeling, Kyiv, Ukraine: Intertechnology, 2009.

K. Sangwook, Y. Zhibin, M. Rhee, and L. Minho, “Deep learning of support vector machines with class probability output networks”, Neural Networks, vol. 64, no. C, pp.19-28, 2015, doi: https://doi.org/​10.1016/j.neunet.2014.09.007.

D. Guthrie, “Unsupervised detection of anomalous text”, Doctor of Philosophy, Department of Computer Science, University of Sheffield, Sheffield, England, 2008.

S. Baccianella, A. Esuli, and F.Sebastiani, “Sentiwordnet 3.0: An enhanced lexical resource for sentiment analysis and opinion mining”, in Proceedings of LREC, 2010, pp. 2200-2204.

P. Stone, “A computer approach to content analysis: Studies using the general inquirer system”, in Spring Joint Computer Conference, New York, 1963, pp. 241-256.

Text classification using neural networks and TensorFlow. [Online]. Available: https://tproger.ru/translations/text-classificationtensorflow-neural-networks. Accessed on: July 17, 2019.

A.Karpathy, “The Unreasonable Effectiveness of Recurrent Neural Networks”, Andrej Karpathy blog, 2015. [Online]. Available: http://karpathy.github.io/2015/05/21/rnn-effectiveness/. Accessed on: April 12, 2019.

I. Witten, E. Frank, and M. Hall, Data Mining: Practical Machine Learning Tools and Techniques 3rd Edition. Waltham, USA: Morgan Kaufmann, 2011.

K. Nigam, “Using maximum entropy for text classification”, in IJCAI99 workshop on machine learning for information filtering, 1999, pp. 61-67.

R. Byrd, S. Hansen, J. Nocedal, and Y. Singer “A Stochastic Quasi-Newton Method for Large-Scale Optimization”, SIAM Journal on Optimization, vol. 26, no. 2, pp. 1008-1031, 2016, doi: https://doi.org/10.1137/140954362.

Published

2020-07-09

How to Cite

Pokrovska, V. (2020). Analysis of information-psychological impact detection methods in social networks. Collection "Information Technology and Security", 8(1), 40–48. https://doi.org/10.20535/2411-1031.2020.8.1.218002

Issue

Section

INFORMATION WARFARE