Fake news detection on social media via implicit crowd signals

The proliferation of Fake News on social media has been a source of widespread concern. One of the main approaches to automatically detect this type of news is based on crowd signals, i.e., opinions manifested by social media users concerning whether the news are fake or not. Although promising, this approach depends on information that is not always available: the explicit opinion of the users about the news to be checked. To overcome this drawback, this article proposes a crowd signal-based method that does not demand the users' explicit opinion to detect Fake News. The proposed method infers the users' opinions from their news spreading (publication/propagation) behavior. Preliminary experiments with two real-world datasets provided evidence that the proposed method can detect Fake News without demanding the explicit opinion of the users about the news and without compromising the classification results obtained by the state-of-the-art crowd signal-based method.