Versini et al. (2020) describe a supervised framework for profiling individuals who disseminate fake news on social media, constructing user representations from four feature families: stylometric cues, inferred personality traits, emotional markers, and semantic embeddings. The system aims to discriminate between authors with a history of sharing false content and those who have not, leveraging writing‑style analysis, behavioral modeling, affective signals, and topic similarity to support authorship profiling in a misinformation‑detection context.
Bears on: authorship & stylometry · concepts & methods