Font: Universitat de Barcelona

Professor convidat: Paolo Rosso, Departament de Sistemes Informàtics i Computació de la Universitat Politècnica de València http://users.dsic.upv.es/~prosso/

Dia: 11 d’abril

Lloc: Edifici Carner (Aribau, 2) aula 0.2

Hora: 17-18

Conferència: Author profiling in social media
Atribución de perfiles en la competición PAN (Uncovering Plagiarism, Authorship, and Wikipedia flaws-2013)

Author profiling is concerned with predicting an author’s demographics from her writing. Besides being personally identifiable, an author’s style may also reveal her age and gender. Accurate predictors are of key interest to forensic linguists and marketers alike.  At PAN-2013 a task is organised and participants are provided with a training data set that consists of documents written in both English and Spanish: http://pan.webis.de/
With regard to the age, posts of three classes are considered: 10s (13-17), 20s (23-27), and 30s (33-47). Moreover, documents from authors who gave a fake age and pretend to be minors will be included (e.g., documents composed of chat lines of sexual predators will be also considered). In this talk, I will give an overview of the state-of-the-art in author profiling, from the identification of age and gender towards native language and personality.