The use of Data Augmentation as a technique for improving neural network accuracy in detecting fake news about COVID-19
It addresses the problem of automated fake news verification for Portuguese speakers, but is incremental as it applies existing techniques to a specific domain.
The paper tackled fake news detection in Portuguese about COVID-19 by applying NLP and data augmentation to a neural network, resulting in significant performance improvement in classification accuracy.
This paper aims to present how the application of Natural Language Processing (NLP) and data augmentation techniques can improve the performance of a neural network for better detection of fake news in the Portuguese language. Fake news is one of the main controversies during the growth of the internet in the last decade. Verifying what is fact and what is false has proven to be a difficult task, while the dissemination of false news is much faster, which leads to the need for the creation of tools that, automated, assist in the process of verification of what is fact and what is false. In order to bring a solution, an experiment was developed with neural network using news, real and fake, which were never seen by artificial intelligence (AI). There was a significant performance in the news classification after the application of the mentioned techniques.