Home // PATTERNS 2023, The Fifteenth International Conference on Pervasive Patterns and Applications // View article
Authors:
Isaac Caicedo-Castro
Mario Macea-Anaya
Samir Castaño-Rivera
Keywords: Machine learning; educational data mining; classification algorithm; dropout and failure forecasting; student long-term retention.
Abstract:
In this work, we ponder the following research question: Is it possible to predict if a given student might either fail or drop out of an undergraduate course taking into account its performance in prerequisite courses? Therefore, we study the case of forecasting the risk faced by students of failing or dropping out of the course of numerical methods in an engineering bachelor's program. To this end, the prediction is based on the student's academic history, which consists of the grades the student has obtained in previous prerequisite courses, whose concepts and skills are required to succeed in the studied case of the numerical methods course. Additionally, the admission test results are also used for forecasting purposes. Moreover, we adopt machine learning, where supervised methods for classification are fitted using the academic history of students enrolled in the Engineering bachelor's program with a major in Systems Engineering at the University of Córdoba in Colombia. We collected the academic history of 56 anonymized students and carried out 10-fold cross-validation. The results of this study reveal that a support vector machines method predicts if a given student is at risk of failing or withdrawing from the numerical methods course with mean values for accuracy, precision, recall, and harmonic mean (F1) of 76.67%, 71.67%, 51.67%, and 57.67%, respectively. This method outperforms the others studied in this work.
Pages: 40 to 51
Copyright: Copyright (c) IARIA, 2023
Publication date: June 26, 2023
Published in: conference
ISSN: 2308-3557
ISBN: 978-1-68558-049-0
Location: Nice, France
Dates: from June 26, 2023 to June 30, 2023