Manipulative Style Recognition of Czech News Texts using Stylometric Text Analysis

SABOL, Radoslav and Aleš HORÁK. Manipulative Style Recognition of Czech News Texts using Stylometric Text Analysis. In Aleš Horák, Pavel Rychlý, Adam Rambousek. Proceedings of the Sixteenth Workshop on Recent Advances in Slavonic Natural Languages Processing, RASLAN 2022. Brno: Tribun EU, 2022, p. 189-197. ISBN 978-80-263-1752-4.

Basic information

• Original name: Manipulative Style Recognition of Czech News Texts using Stylometric Text Analysis
• Authors: SABOL, Radoslav (703 Slovakia, guarantor, belonging to the institution) and Aleš HORÁK (203 Czech Republic, belonging to the institution).
• Edition: Brno, Proceedings of the Sixteenth Workshop on Recent Advances in Slavonic Natural Languages Processing, RASLAN 2022, p. 189-197, 9 pp. 2022.
• Publisher: Tribun EU

Other information

• Original language: English
• Type of outcome: Proceedings paper
• Field of Study: 10200 1.2 Computer and information sciences
• Country of publisher: Czech Republic
• Confidentiality degree: is not subject to a state or trade secret
• Publication form: printed version "print"
• WWW: Domovská stránka workshopu Plný text
• RIV identification code: RIV/00216224:14330/22:00127486
• Organization unit: Faculty of Informatics
• ISBN: 978-80-263-1752-4
• ISSN: 2336-4289
• Keywords in English: stylometry; propaganda detection; manipulative style analysis; Propaganda dataset; Czech

Abstract

The rampant spread of manipulative texts purporting propaganda, disinformation or surveillance, requires the readers to take heed of the actual reasoning behind and the real purpose of the newspaper texts. The capability to recognize a malignant content asks for more and more concentration and background knowledge. A support offered by automated content analysis tools forms an important part of such protective approaches. In the presented text, we introduce a new approach to detecting a set of manipulative stylistic techniques in Czech newspaper texts by exploiting stylometric methods in conjunction with deep learning text classification. We show that the stylometric analysis with almost 20,000 features allows to improve the results for most of the techniques. The results are evaluated with a previously published Czech Propaganda dataset.

Links

• LM2018101, research and development project: Name: Digitální výzkumná infrastruktura pro jazykové technologie, umění a humanitní vědy (Acronym: LINDAT/CLARIAH-CZ)

Investor: Ministry of Education, Youth and Sports of the CR