Through a systematic literature review method, in this work we searched classical electronic libraries in order to find the most recent papers related to fake news detection on social medias. Our target is mapping the state of art of fake news detection, defining fake news and finding the most useful machine learning technique for doing so. We concluded that the most used method for automatic fake news detection is not just one classical machine learning technique, but instead a amalgamation of classic techniques coordinated by a neural network. We also identified a need for a domain ontology that would unify the different terminology and definitions of the fake news domain. This lack of consensual information may mislead opinions and conclusions.

No abstract

Nowadays it is hard to distinguish what is a fake information made to mislead, cause havoc and hysteria than a true critic that has only the intention to highlight a social problem or an abnormality of some sort. In order to diminish the number of false positives in fake news detection, we experimented two neural network models trained by a combined set of true, fake and sarcastic news in order to test how accurate would be our model. This paper has the goal to propose future steps of an ongoing experiment and discuss the usage of collaboration aided by machines to gather such data to the models.

No abstract

Information sharing on the Web has also led to the rise and spread of fake news. Considering that fake information is generally written to trigger stronger feelings from the readers than simple facts, sentiment analysis has been widely used to detect fake news. Nevertheless, sarcasm, irony, and even jokes use similarwritten styles, making the distinction between fake and fact harder to catch automatically. We propose a new fake news Classifier that considers a set of language attributes and the gradient of sentiments contained in a message. Sentiment analysis approaches are based on labelling news with a unique value that shrinks the entire message to a single feeling. We take a broader view of a message’s sentiment representation, trying to unravel the gradient of sentiments a message may bring. We tested our approach using two datasets containing texts written in Portuguese: a public one and another we created with more up-to-date news scrapped from the Internet. Although we believe our approach is general, we tested for the Portuguese language. Our results show that the sentiment gradient positively impacts the fake news classification performance with statistical significance. The F-Measure reached 94 %, with our approach surpassing available ones (with a p-value less than 0.05 for our results).