Análise de sentimentos em postagens de redes sociais relacionadas à comunidade LGBTQIA+: desafios, técnicas e contribuições

Abstract

This thesis examined different sentiment classification methods in social media posts related to the LGBTQIA+ community. The VADER, Random Forest, CNN, and RNN methods were implemented and evaluated. The results highlight the importance of sentiment analysis in this context, with Random Forest achieving the highest accuracy. Sentiment analysis can provide an overview of the emotions and opinions expressed within the LGBTQIA+ community on social media, contributing to understanding their perception and emotional impact. Future research should focus on optimizing the methods and consider ethical aspects to ensure more accurate and inclusive results.