Deciding among fake, satirical, objective and legitimate news: A multi-label classification system

Currently, the widespread of fake news has raised on the politicalclass and society members in general, increasing concerns aboutthe potential of misinformation that can be propagated, appearingon the center of the debate about election results around the world.On the other hand, satirical news has an entertaining purpose andare mistakenly put on the same boat of objective fake news. Inthis work, we address the differences between objectivity and legitimacy of news documents, treating each article as having twoconceptual classes: objective/satirical and legitimate/fake. Thus, wepropose a Decision Support System (DSS) based on a text miningpipeline and a set of novel textual features that uses multi-labelmethods for classifying news articles on those two domains. Forvalidating the approach, a set of multi-label methods was evaluatedwith a combination of different base classifiers and then comparedto a multi-class approach. Results reported our DSS as proper (0.80F1-score) in addressing the scenario of misleading news from challenging perspective of multi-label modeling, outperforming themulti-class methods (0.71 F1-score) over a real-life news datasetcollected from several portals of news.