Multi-dimensional feature extraction for the identification of online harm spreaders

Abstract

Identifying online harm spreaders, such as fake news or hate speech spreaders, has become an important problem to prevent the propagation of low-quality content online. In this work, several dimensions of the user behaviour are defined to analyze this type of users and a set of features are extracted within each dimension, leveraging mostly lexicons, in order to better characterize spreaders and help to its detection.

Publication
Proceedings of the Latin American Workshop on Information Fusion (LAFUSION 2023)