La subjectivité dans le journalisme québécois et belge : Transfert de connaissances inter-médias et inter-cultures

Escouflaire, Louis;Venant, Antoine;Descampe, Antonin;Fairon, Cédrick
(2024) JADT 2024: 17th International Conference on Statistical Analysis of Textual Data — Location: Brussels, Belgium (25.June.2024)

Files

jadt_belqbc_final_v4.pdf
  • Open Access
  • Adobe PDF
  • 1.33 MB
Download

Details

Authors
Abstract
We compare the ability to transfer knowledge across journalistic sources and cultures for an opinion vs. news articles classification task in French, between two models: the large language model CamemBERT and a statistical model based on linguistic features. We use a corpus of 80,000 articles published by eight Canadian and Belgian French media, which also allows us to compare how opinion and news articles contrast in the two media landscapes. Our results show that CamemBERT has a higher knowledge transfer capacity than the feature-based model, and attest to the existence of a difference between Quebec and Belgian press discourse.
Affiliations

Citations

  • APA
  • Chicago
  • FWB

Escouflaire, L., Venant, A., Descampe, A., & Fairon, C. (2024). La subjectivité dans le journalisme québécois et belge : Transfert de connaissances inter-médias et inter-cultures. JADT 2024 Proceedings: 17th International Conference on Statistical Analysis of Textual Data. 2022. Volume 1., 329-338. https://hdl.handle.net/2078.5/269261