Abstract
We propose a novel Rhetoric Mining methodology to identify moves of persuasion used in disinformation of social media news posts. Rhetoric Mining combines qualitative methodologies and rhetorical theory analysis with machine learning techniques to automatically identify rhetorical moves. Rhetorical moves are instances of discourse intentionally used to persuade an audience. Rhetoric Mining converts the qualitative detection of persuasive moves into quantified rhetoric instance vectors which can be used to characterize rhetorical styles of a text. We identify rhetorical styles of persuasion (news posts with high positive responses, likes, shares, or re-posts) as well as disinformation (news posts that are persuasive but false).
Recommended Citation
Seref, Michelle MH and Seref, Onur, "Rhetoric Mining for Fake News: Identifying Moves of Persuasion and Disinformation" (2019). AMCIS 2019 Proceedings. 1.
https://aisel.aisnet.org/amcis2019/rhetoric_social_media_disinformation/rhetoric_social_media_disinformation/1
Rhetoric Mining for Fake News: Identifying Moves of Persuasion and Disinformation
We propose a novel Rhetoric Mining methodology to identify moves of persuasion used in disinformation of social media news posts. Rhetoric Mining combines qualitative methodologies and rhetorical theory analysis with machine learning techniques to automatically identify rhetorical moves. Rhetorical moves are instances of discourse intentionally used to persuade an audience. Rhetoric Mining converts the qualitative detection of persuasive moves into quantified rhetoric instance vectors which can be used to characterize rhetorical styles of a text. We identify rhetorical styles of persuasion (news posts with high positive responses, likes, shares, or re-posts) as well as disinformation (news posts that are persuasive but false).