Guia comparativo de classificadores de texto para os ODS

Abstract

Text classification algorithms are computational tools that automate the task of assigning categories or labels to texts based on their content. The aim of this study is to conduct a systematic literature review to identify and analyze the main classification algorithms used in categorizing materials related to the 17 Sustainable Development Goals (SDGs). The analysis of the database revealed a significant gap in the availability of resources dedicated to this theme, suggesting that this is still a largely unexplored field, especially considering the deadline for the conclusion and evaluation of the Agenda 2030 results. The main classification algorithms used in this context have been identified, with Logistic Regression, Naive Bayes, Support Vector Machine, Random Forest, Neural Networks, K-Nearest Neighbor, and Decision Tree standing out. Among them, the Support Vector Machine demonstrated consistent performance and achieved excellent classification results, being widely addressed in the reviewed studies. This research contributes to understanding the main classification algorithms and identifying the most suitable one for such analyses, aiming to enhance the achievement of the goals established by the SDGs.