MINERAÇÃO DE TEXTOS NA REDE SOCIAL X: Uma análise comparativa de algoritmos de classificação e técnicas de processamento de linguagem natural

Abstract

This article conducts a comparative analysis of classification algorithms and natural language processing (NLP) techniques for text mining on social network X (formerly Twitter). The study aims to identify the most effective combination of classification algorithm and NLP techniques, employing methods like Bag of Words, TF-IDF, lemmatization, and stemming, in conjunction with Logistic Regression, Naive Bayes, and Support Vector Machine (SVM) algorithms. Through rigorous methodology, the research compares the efficacies of these techniques and algorithms, offering practical insights applicable in various fields ranging from digital marketing to sentiment analysis and public opinion research. The findings highlight the superiority of the Logistic Regression model, particularly when combined with the Bag of Words technique and lemmatization approach, proving to be a robust strategy for sentiment classification in tweets across multiple languages.