Pasar al contenido principal

Resultados relacionados

Resultados totales (Incluyendo duplicados): 1
Encontrada(s) 1 página(s)

RUA. Repositorio Institucional de la Universidad de Alicante
Publicación Otros (Other)

Semi-automatic Annotation Proposal for Increasing a Fake News Dataset in Spanish

  • Bonet-Jover, Alba
  • Bonet-Jover, Alba|||/items/6bd6e7a1-46a7-415c-bdde-2a1e56f97a2f
The digital era has become an ally of fake news, since it has increased the spread and amount of false information. Fake news is a global problem that causes disorder and generates fear. This phenomenon must be attacked in the same environment in which it is generated: in the digital environment. This paper presents the current state of my doctoral thesis which focuses on the linguistic modelling applied to the automatic detection of fake news through Natural Language Processing (NLP). In order to study the linguistic characteristics of fake news and to create computational models that automate its detection, labelled datasets are needed, but this is a costly task that requires time and expertise. A fake news dataset and an annotation guide were created ad hoc in a previous work to analyse all the parts and elements of a news item. However, after creating and training our system, we realised that the time spent was not proportional to the low annotated data obtained. The need of creating a larger corpus to train and test our hypothesis has led us to think about a way of increasing our corpus without spending so much time. For that purpose, a semi-automatic annotation is proposed for reducing time while increasing speed and quantity of the examples annotated. This proposal, besides allowing us to make progress in our research, may facilitate the creation of datasets, which are essential in NLP research.