Abstract
As Brazil faced one of its most important elections in recent times, the fact-checking agencies handled the same kind of misinformation that has attacked voting in the US. However, stopping fake content before it goes viral remains an intense challenge. This paper examines a sample database of the 2018 Brazilian election articles shared by Brazilians over social media platforms. We evaluated three different configuration of Long Short-Term Memory. Experiment results indicate that the 3-layer Deep BiLSTMs with trainable word embeddings configuration was the best structure for fake news detection. We noticed that the developments in deep learning could potentially benefit fake news research.
Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
