Application of Logistic Regression to Predict the Failure of Students in Subjects of a Mathematics Undergraduate Course

Autor/inn/en	Costa, Stella F.; Diniz, Michael M.
Titel	Application of Logistic Regression to Predict the Failure of Students in Subjects of a Mathematics Undergraduate Course
Quelle	In: Education and Information Technologies, 27 (2022) 9, S.12381-12397 (17 Seiten)Infoseite zur Zeitschrift PDF als Volltext Verfügbarkeit
Zusatzinformation	ORCID (Costa, Stella F.)
Sprache	englisch
Dokumenttyp	gedruckt; online; Zeitschriftenaufsatz
ISSN	1360-2357
DOI	10.1007/s10639-022-11117-1
Schlagwörter	Prediction; Undergraduate Students; Mathematics Education; Models; Information Retrieval; Pattern Recognition; Data Analysis; Information Technology; Foreign Countries; Academic Failure; Predictor Variables; Brazil + Suchen Sie Ihr Suchwort? Vorhersage; Mathematische Bildung; Analogiemodell; Mustererkennung; Auswertung; Informationstechnologie; Ausland; Prädiktor; Brasilien
Abstract	The large rates of students' failure is a very frequent problem in undergraduate courses, being even more evident in exact sciences. Pointing out the reasons of such problem is a paramount research topic, though not an easy task. An alternative is to use Educational Data Mining techniques (EDM), which enables one to convert data from educational database into useful information, in order to understand and improve teaching and learning processes. In this way, the objective of this paper is to propose mathematical models based on EDM techniques to estimate the probability of a student in a mathematics degree course at IFSP (Federal Institute of São Paulo) to fail in exact sciences disciplines, and later on, indicate which aspects contribute significantly for the Students' failure rates in these branches. We present three logistic regression models that which were applied based on socioeconomic data and student performance over 4 years. For interpretation and evaluation of such models, odds ratio, ten-fold Cross Validation method and the metrics: accuracy, sensitivity, specificity and area under the ROC curve (AUC) were used. It was noted that through Cross Validation, the models achieved accuracy values accounting for over 70%, sensitivity over 70%, specificity over 60% and AUC over 0.75. Analyzing the predictive variables of these models, we identified that factors such as advantage age, rates of failure through the course and attendance in initial semesters can increase the probability of failure in exact science disciplines in the analyzed course. (As Provided).
Anmerkungen	Springer. Available from: Springer Nature. One New York Plaza, Suite 4600, New York, NY 10004. Tel: 800-777-4643; Tel: 212-460-1500; Fax: 212-460-1700; e-mail: customerservice@springernature.com; Web site: https://link.springer.com/
Erfasst von	ERIC (Education Resources Information Center), Washington, DC
Update	2024/1/01