Original scientific paper
https://doi.org/10.1080/00051144.2023.2296790
Effective deep learning based grade prediction system using gated recurrent unit (GRU) with feature optimization using analysis of variance (ANOVA)
S. Lakshmi
; Department of Computer Science and Engineering, Noorul Islam Centre for Higher Education, Kumaracoil, India
*
C. P. Maheswaran
; Department of Artificial Intelligence and Data Science, Sri Krishna College of Technology, Coimbatore, India
* Corresponding author.
Abstract
The prediction of final semester grades is a crucial undertaking in education, offering insights
into student performance and enabling timely interventions to support their academic journey.
This paper employs a deep learning approach, specifically gated recurrent unit (GRU), in conjunction with feature optimization using analysis of variance (ANOVA), to forecast final semester
grades. The predictive model is trained and evaluated on a handcrafted grade prediction dataset,
which contains the academic performance of the students during Plus 2 and from Semester 1
to Semester 4 of a group of Computer science and Engineering majors in Kerala. By processing historical academic records and contextual information, the GRU model learns to predict
future performance accurately. To enhance the model’s efficacy and interpretability, ANOVA is
applied to optimize the feature selection process. This statistical technique identifies the most
influential factors contributing to final grades, refining the model’s predictive power while reducing dimensionality. The experimental results showcase the model’s effectiveness in predicting
final semester grades, demonstrating superior accuracy and performance compared to grade
prediction using CNN with Bayesian optimization and LSTM with L1-Norm optimization.
Keywords
Student performance prediction; artificial intelligence; deep learning; recurrent neural network; gated recurrent unit; feature optimization; ANOVA
Hrčak ID:
323035
URI
Publication date:
10.1.2024.
Visits: 0 *