Open AccessPredicting Dropout in Higher Education: a Systematic Review
Author(s) -
Jailma Januário da Silva,
Norton Trevisan Roman
Publication year - 2021
Publication title -
anais do xxxii simpósio brasileiro de informática na educação (sbie 2021)
Language(s) - English
Resource type - Conference proceedings
DOI - 10.5753/sbie.2021.217437
Subject(s) - computer science , metric (unit) , dropout (neural networks) , decision tree , precision and recall , inclusion (mineral) , recall , machine learning , theme (computing) , systematic review , protocol (science) , artificial intelligence , data mining , information retrieval , data science , world wide web , engineering , psychology , medicine , social psychology , operations management , alternative medicine , medline , pathology , law , political science , cognitive psychology
In this article, we present a systematic literature review, carried out from February to March 2020, on the application of a machine learning technique to predict student dropout in higher education institutions. Besides describing the protocol followed during our research, which includes the research questions, searched databases and query strings, along with criteria for inclusion and exclusion of articles, we also present our main results, in terms of the attributes used by current research on this theme, along with adopted approaches, specific algorithms, and evalution metrics. The Decision Tree technique is the most used for the construction of models, and accuracy and recall and precision being the most used metric for evaluating models.