Early Prediction of Student Performance Using Deep Neural Networks

Authors

  • P. Sankara Rao, V. Sai Snehitha, P. N. D. Yamini, P. Jagadeesh, T. Aakash

DOI:

https://doi.org/10.17762/msea.v71i4.764

Abstract

In contemporary educational systems, the ability to predict student performance is deteriorating gradually. Predicting a student's performance in advance can help students and teachers track a student's development. In today's world, a continuous evaluation strategy has been adopted by a great number of institutions. These kinds of systems are beneficial to students because they improve their overall academic performance. The encouragement of regular students is the goal of the continuous evaluation process. In recent years, neural networks have been implemented successfully in a variety of data mining applications, frequently outperforming traditional classifiers. In this paper, we used deep Neural Networks to predict student performance using Learning Management System (LMS) data within the context of Educational Data Mining. The various training features are derived from LMS data collected throughout the duration of each course and include usage statistics such as class test, assignment, attendance, and mid-exam marks in order to determine which of these factors are associated with the students' overall performance. The proposed method extracts informative data weighted appropriately. In this technique for generating neural networks, numerous updated hidden layers are utilized. These layers are determined based on feed forwarding and back propagation data from previous cases. Precision, recall, F1 score, accuracy and root mean squared error are used to evaluate the proposed method performance. On the basis of these results, we can conclude that deep neural networks outperform all existing state-of-the-art methods and could be used to accurately predict future student performance.

Downloads

Published

2022-09-09

How to Cite

P. Sankara Rao, V. Sai Snehitha, P. N. D. Yamini, P. Jagadeesh, T. Aakash. (2022). Early Prediction of Student Performance Using Deep Neural Networks. Mathematical Statistician and Engineering Applications, 71(4), 2130–2143. https://doi.org/10.17762/msea.v71i4.764

Issue

Section

Articles