Показати скорочений опис матеріалу
dc.contributor.author | Volkova, Nataliia![]() |
|
dc.date.accessioned | 2024-08-26T11:48:09Z | |
dc.date.available | 2024-08-26T11:48:09Z | |
dc.date.issued | 2024 | |
dc.identifier.citation | Volkova Nataliia. Computational Detection and Analysis of Manipulation Techniques in News Channels on Telegram in Ukraine. Ukrainian Catholic University, Faculty of Applied Sciences, Department of Computer Sciences. Lviv 2024, 36 p. | uk |
dc.identifier.uri | https://er.ucu.edu.ua/handle/1/4681 | |
dc.language.iso | en | uk |
dc.subject | Computational Detection | uk |
dc.subject | Analysis of Manipulation Techniques | uk |
dc.subject | News Channels on Telegram | uk |
dc.subject | Ukraine | uk |
dc.subject | Telegram | uk |
dc.title | Computational Detection and Analysis of Manipulation Techniques in News Channels on Telegram in Ukraine | uk |
dc.type | Preprint | uk |
dc.status | Публікується вперше | uk |
dc.description.abstracten | In this research, we aim to identify specific instances of manipulation techniques such as doubts, black-and-white fallacy, appeal to fear, and loaded language and their granularity in news text from social media. To achieve this, we developed our own manually annotated corpus with 1,877 posts from Ukrainian news chan- nels on Telegram, all in the Ukrainian language, consisting of 3,472 manipulation techniques. Each annotation includes the fragment or span where a manipulation technique is detected, along with the corresponding technique from a set of selected techniques. We then trained a pre-trained BERT model to recognize these spans and their associated manipulation techniques. Additionally, we generated syntactic data to enhance the model’s performance. | uk |