Lecture applies Natural Language Processing to identify Fake News

Recognition programs analyze the most frequently occurring words and structures in fake news.

On May 24, 2022, ESEG – Faculdade do Grupo Etapa held a lecture on the application of Natural Language Processing in the recognition of Fake News, given by Professor Alexandre Albano Pupo Vizotto, Master in Economics and with an MBA in Data Science and Decision. A segment of Artificial Intelligence, NLP is dedicated to technologically understanding human language.

Throughout the lecture, the professor introduced basic concepts such as Corpus and Metadata, DTM and tidytext, as well as briefly presenting different systems used in NLP, such as Bag of Words (which takes into account the words used in a given sample) and Semantic Parsing (which analyzes the semantic structure of the sample text). Based on these concepts, the professor complemented his explanation by conducting a live language processing exercise for the students.

During the demonstration, Vizotto presented processes for applying formulas that recognize which words are most frequently used in fake news and in real news, based on a pre-determined sample. With the information obtained, the professor showed that it is possible to analyze the content of these fake news stories. He commented, however, that programs like this are not yet 100% accurate, especially due to the lack of Portuguese dictionaries used for recognizing these words. The development of these dictionaries in Brazil is still recent, and therefore, there are few available.

The meeting was marked by student participation with questions and comments.