Computational Intelligence Enabled Student Performance Estimation in the Age of COVID-19

Area of Science:

• Artificial Intelligence
• Educational Technology
• Data Science

Background:

• The COVID-19 pandemic disrupted traditional student evaluations, highlighting the need for flexible assessment methods.
• Difficulties in online participation necessitate alternative approaches for student performance estimation.

Purpose of the Study:

• To investigate the efficacy of deep learning (DL) and machine learning (ML) for predicting student grades.
• To develop an automated system for estimating student performance using incomplete academic data.

Main Methods:

• Analysis of DL and ML algorithms for student performance prediction.
• Development of a fully data-driven estimation system, avoiding manual feature engineering.
• Utilizing a publicly available dataset comprising 15 courses for academic research.

Main Results:

• Latent space models demonstrated superior performance compared to sequential models.
• Deep learning models showed high accuracy in estimating student performance.
• Prediction accuracy increased with larger training datasets.

Conclusions:

• Deep learning offers a robust and generic solution for automated student performance estimation.
• A greater number of low-weightage assessments are more effective for accurate grade prediction than fewer high-weightage exams.
• The data-driven DL approach provides a scalable and adaptable solution for educational assessment.