.png){kind=icon}
.png){kind=icon}
Return
A transformer-enhanced framework for estimating student performance in mathematics
.png){kind=icon}
PRE
AI
W
Z
DOI：10.7717/peerj-cs.3692.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Estimating student performance is a crucial focus in educational research, significantly impacting academic outcomes and instructional strategies. This article investigates the application of advanced deep learning techniques for predicting student performance based on heterogeneous tabular data. We propose a novel hybrid architecture that combines a Hybrid Transformer module with a Multi-Layer Perceptron (MLP) layer, designed to effectively capture both static and dynamic feature interactions. Motivated by recent advancements in Transformer-based models for tabular data, our approach integrates self-attention mechanisms with feed-forward pathways to enhance feature learning. Experimental results demonstrate that the proposed model outperforms traditional machine learning methods and state-of-the-art tabular deep learning architectures across multiple evaluation metrics. These findings highlight the potential of hybrid Transformer-MLP architectures in optimizing predictive analytics for educational outcomes.
Keywords:
Student performance prediction
Transformer-based deep learning
Educational data mining
Tabular data modeling
Hybrid neural network architecture
Journal
IF:
2.5
Papers:
3.3K
Citations:
6.9K
Organization
Cited Papers