Identificación automática de noticias falsas en español utilizando técnicas de minería de datos y procesamiento del lenguaje natural.

Abstract

The circulation of false news, especially political satire type, via digital media has affected the majority of the Ecuadorian population due to its massive spread and lack of regulation, which is why in this paper presents a methodology grounded in statistical learning that accurately and automatically detects fake news based on the way it is written. For this, natural processing language, classical supervised classification algorithms and the Python programming language were used. The document begins by defining all the necessary concepts to understand both false news and statistical learning algorithms, later it is explained how the news was collected and processed, to then detail the process of training the models and finally the presentation and analysis of the respective results.