Logistic Regression on Guard of Students’ Academic Performance

Artificial Intelligence and System Engineering: Proceedings of 8th Computational Methods in Systems and Software 2024. Vol.2. Lecture Notes in Networks and Systems. Vol.1490 2025
Hoang Phuong Nguyen, Ivana Roncevic, Elena Denisova, Maria Anatolyevna Ivanova, Igor Fedorchenko, Nataļja Muračova, Roman Tsarev

Learning is an integral and often labor-intensive part of any person’s life activity. University education is primarily aimed at training a highly qualified specialist. However, due to a number of reasons, students may have difficulties in learning, which can lead to expulsion from the university. In today’s educational landscape, it becomes an essential issue to identify students who are at such a risk zone in order to provide them with the necessary and timely support. In this paper, we consider the application of logistic regression to predict student dropout at early stages. Based on the statistics collected, a model is developed that allows after some time from the beginning of the semester to calculate the probability that a student will be expelled. This indicator allows the teacher to pay special attention to the underperforming students and direct his efforts to increase their motivation, provide them with suitable study material and other assistance to avoid expulsion of these students due to their failure.

Keywords
E-Learning; Logit Transformation; Multiple Regression; Predictor; Probability; Regression; Regressor
DOI
10.1007/978-3-031-96759-7_26
Hyperlink
https://link.springer.com/chapter/10.1007/978-3-031-96759-7_26

Nguyen, H., Roncevic, I., Denisova, E., Ivanova, M., Fedorchenko, I., Muračova, N., Tsarev, R. Logistic Regression on Guard of Students’ Academic Performance. In: Artificial Intelligence and System Engineering: Proceedings of 8th Computational Methods in Systems and Software 2024. Vol.2. Lecture Notes in Networks and Systems. Vol.1490, Russia, Moscow, 12-14 October, 2024. Cham: Springer, 2025, pp.366-374. ISBN 978-3-031-96758-0. e-ISBN 978-3-031-96759-7. ISSN 2367-3370. e-ISSN 2367-3389. Available from: doi:10.1007/978-3-031-96759-7_26

Publication language
English (en)