Abstract
The problem of determining the author of a text, formerly solved on fiction texts, has currently become one of the central ones in cybersecurity, which calls for new techniques and approaches to be employed. One of the issues in the field is a short length of texts whose authorship is typically needed to be determined and a small amount of data per author. While a few papers investigate the problem of the minimal text length for authorship analysis, the effect of the total amount of data per author on the accuracy of attribution is not well studied, especially in regards to informal texts. The paper analyzes the effect of the total volume of texts per author on the idiolect identification accuracy separately for different types of stylometric markers on the datasets compiled of Russian forum texts. For text classification a supervised version of principal component analysis (PLS-DA) is applied which is widely used in bioinformatics, but it has been introduced for the first time for an idiolect identification task. The conclusion is made about the absence of a linear effect of the total amount of data on the attribution accuracy. It is assumed that author recognizability for a classifier depends on a variety of factors which are unique to each particular attribution case.
This is a preview of subscription content, log in via an institution to check access.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Similar content being viewed by others
Notes
- 1.
TreeTagger - a part-of-speech tagger for many languages. https://www.cis.uni-muenchen.de/~schmid/tools/TreeTagger/ (assessed on March, 7 2021).
References
Burrows, J.: All the way through: testing for authorship in different frequency strata. Lit. Linguist. Comput. 22(1), 27–47 (2007)
Demsar, J.: Statistical comparisons of classifiers over multiple data sets. J. Mach. Learn. Res. 7(Jan), 1–30 (2006)
Eder, M.: Does size matter? Authorship attribution, small samples, big problem. Digital Scholarsh. Hum. 30(2), 167–182 (2015)
Eder, M.: Short samples in authorship attribution: a new approach. In: Digital Humanities 2017: Conference Abstracts, pp. 221–24. McGill University, Montreal (2017)
Gorman, R.: Author identification of short texts using dependency treebanks without vocabulary. Digital Scholarsh. Hum. 35(4), 812–825 (2020)
Kapočiūtė-Dzikienė, J., Šarkutė, L., Utka, A.: The effect of author set size in authorship attribution for Lithuanian. In: Proceedings of the 20th Nordic Conference of Computational Linguistics (NODALIDA 2015), pp. 87–96. Vilnius, Lithuania (2015)
Koppel, M., Schler, J., Argamon, S.: Authorship attribution in the wild. Lang. Resour. Eval. 45(1), 83–94 (2011)
Koppel, M., Schler, J., Argamon, S., Messeri, E.: Authorship attribution with thousands of candidate authors. In: Proceedings of the 29th Annual International ACM SIGIR Conference on Research and Development in Information Retrieval (SIGIR 2006), pp. 659–660. Association for Computing Machinery, New York, NY, USA (2006)
Kuhn, M.: Caret: Classification and Regression Training. R package version 6.0–73. Vienna, Austria: R Foundation for Statistical Computing (2016)
Litvinova, T.: Process-oriented characteristics of an idiolect for authorship attribution of heterogeneous texts: a pilot study. In: Proceedings of the Computational Models in Language and Speech Workshop (CMLS 2020) co-located with 16th International Conference on Computational and Cognitive Linguistics (TEL 2020), pp. 3–17. CEUR-WS.org, Kazan (2020)
Litvinova, T.: Rusidiolect: a new resource for authorship studies. In: Antipova, T. (ed.) ICCS 2020. LNNS, vol. 186, pp. 14–23. Springer, Cham (2021). https://doi.org/10.1007/978-3-030-66093-2_2
Luyckx, K., Daelemans, W.: The effect of author set size and data size in authorship attribution. Lit. Linguist. Comput. 26(1), 35–55 (2011)
Panicheva, P., Litvinova, T.: Authorship attribution in Russian in real-world forensics scenario. In: Martín-Vide, C., Purver, M., Pollak, S. (eds.) SLSP 2019. LNCS (LNAI), vol. 11816, pp. 299–310. Springer, Cham (2019). https://doi.org/10.1007/978-3-030-31372-2_25
Rohart, F., Gautier, B., Singh, A., Lê Cao, K.-A.: mixOmics: an R package for ‘omics’ feature selection and multiple data integration. PLoS Comput. Biol. 13(11), e1005752 (2017)
Ruiz-Perez, D., Guan, H., Madhivanan, P., Mathee, K., Narasimhan, G.: So you think you can PLS-DA? BMC Bioinform. 21(1), 1–10 (2020)
Acknowledgement
This work is supported by the grant N 18-78-10081 from Russian Science Foundation, which is gratefully acknowledged.
Author information
Authors and Affiliations
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2022 The Author(s), under exclusive license to Springer Nature Switzerland AG
About this paper
Cite this paper
Litvinova, T. (2022). Authorship Attribution of Russian Social Media Texts: Does the Volume of Data Favor Idiolect Identification?. In: Antipova, T. (eds) Comprehensible Science. ICCS 2021. Lecture Notes in Networks and Systems, vol 315. Springer, Cham. https://doi.org/10.1007/978-3-030-85799-8_30
Download citation
DOI: https://doi.org/10.1007/978-3-030-85799-8_30
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-030-85798-1
Online ISBN: 978-3-030-85799-8
eBook Packages: Intelligent Technologies and RoboticsIntelligent Technologies and Robotics (R0)