References & Citations
Computer Science > Computation and Language
Title: Exploring Thematic Coherence in Fake News
(Submitted on 16 Dec 2020 (v1), last revised 17 Dec 2020 (this version, v2))
Abstract: The spread of fake news remains a serious global issue; understanding and curtailing it is paramount. One way of differentiating between deceptive and truthful stories is by analyzing their coherence. This study explores the use of topic models to analyze the coherence of cross-domain news shared online. Experimental results on seven cross-domain datasets demonstrate that fake news shows a greater thematic deviation between its opening sentences and its remainder.
Submission history
From: Martins Samuel Dogo [view email][v1] Wed, 16 Dec 2020 18:01:04 GMT (425kb)
[v2] Thu, 17 Dec 2020 01:56:29 GMT (425kb)
Link back to: arXiv, form interface, contact.