An exploratory analysis on socio-demographic features importance for a predictive undergraduate computing students dropout model

School dropout is a problem faced by educational systems worldwide across various levels of education and institutions. In this regard, several strategies are studied and tested to address this issue or at least mitigate it. With the advancement of artificial intelligence, particularly machine learning, a promising opportunity arises to develop robust predictive models capable of accurately identifying complex patterns and anticipating dropout cases. This study explores the alternatives found by some authors in using machine learning to prevent school dropout, highlighting and comparing aspects of feature engineering adopted and the most relevant characteristics in the training process. Analyzing case studies and recent research, this work demonstrates the most important variables and the ones most chosen among researchers to create machine learning models, suggesting which paths are more efficient and faster for new research.