Logotipo do repositório
 

Publicação:
Detecção de fake news em português utilizando inferência de linguagem natural

Carregando...
Imagem de Miniatura

Orientador

Marana, Aparecido Nilceu

Coorientador

Pós-graduação

Curso de graduação

Bauru - FC - Ciência da Computação

Título da Revista

ISSN da Revista

Título de Volume

Editor

Universidade Estadual Paulista (Unesp)

Tipo

Trabalho de conclusão de curso

Direito de acesso

Acesso abertoAcesso Aberto

Resumo

Resumo (português)

Com o advento da Internet, é possível obter informações a partir de quaisquer fontes sobre eventos ao redor do mundo, inclusive informações imprecisas e até mesmo falsas, sendo estas chamadas de fake news. A divulgação e a proliferação de fake news podem causar prejuízos e danos seríssimos para as pessoas e para a sociedade de modo geral, salientando a extrema importância de sua identificação. Isso pode ser feito com o uso do Processamento de Linguagem Natural, que abrange técnicas computacionais para a análise automática e à representação de linguagens antropológicas. Por exemplo, a detecção de fake news pode ser feita com a Inferência de Linguagem Natural, uma sub-área do Processamento de Linguagem Natural focada na verificação da implicação ou contradição em um par de sentenças, que pode ser utilizada neste contexto considerando notícias já confirmadas como sendo verdadeiras como a premissa e uma notícia suspeita como a hipótese, dessa forma, caso haja contradição entre a hipótese e a premissa, a notícia suspeita é considerada falsa, caso contrário, ela é rotulada como verdadeira. Visto isso, o objetivo deste trabalho é propor um método de detecção automática de fake news para a língua portuguesa utilizando a Inferência de Linguagem Natural. Como não há uma base de dados em português que utilize a Inferência de Linguagem Natural para esta aplicação, é possível traduzir dados já existentes para este idioma, em vista disso, neste trabalho utilizou-se a base de dados em inglês denominada FNID-FakeNewsNet. Com base nos resultados obtidos para o conjunto de teste, observa-se um desempenho melhor quando a premissa é utilizada junto à hipótese. Além disso, de modo geral, é possível observar que os resultados não foram afetados significativamente pela tradução dos dados em inglês para o português.

Resumo (inglês)

With the Internet, it is possible to obtain news about global events regardless of the source, which means that inaccurate and even false information, called fake news, is shared. This dissemination and proliferation of fake news can cause harm and damage to people and society, highlighting the extreme importance of their identification. One way to approach this problem is by using Natural Language Processing, which encompasses computational techniques for the automatic analysis and representation of anthropological languages. For example, the detection of fake news is possible with Natural Language Inference, a sub-area of Natural Language Processing focused on checking the implication or contradiction in a pair of sentences, which can be used in this context considering reliable news as the premise and suspicious news as the hypothesis, thus, if there is a contradiction between them, the investigated news is considered false, otherwise it is labeled as true. Given this, the objective of this work is to propose a method for the automatic detection of fake news in the Portuguese language using Natural Language Inference. As no database in Portuguese uses Natural Language Inference for this application, it is possible to translate existing data into this language. Therefore, in this work, we used the database in English called FNID-FakeNewsNet. Based on the results obtained from the test set, better performance occurs with the simultaneous use of premise and hypothesis. Furthermore, it is also possible to observe that the results were not significantly affected by the translation from English to Portuguese.

Descrição

Palavras-chave

Fake news, Inferência de linguagem natural, Processamento de linguagem natural, Natural language inference, Natural language processing

Idioma

Português

Como citar

Hondo, Larissa Mayumi Barela. Detecção de fake news em português utilizando inferência de linguagem natural. 2023. Trabalho de Conclusão de Curso (Bacharelado em Ciência da Computação) - Universidade Estadual Paulista, Unesp, Bauru, 2023.

Itens relacionados

Financiadores

Unidades

Departamentos

Cursos de graduação

Programas de pós-graduação