Predictive modelling and analytics of students' grades using machine learning algorithms

Area of Science:

• Educational Technology
• Machine Learning in Education
• Student Performance Prediction

Background:

• The COVID-19 pandemic necessitated a shift to online learning, increasing reliance on digital platforms.
• Predicting student academic performance is crucial for timely interventions and support.
• Online learning platforms offer rich interaction data valuable for performance analysis.

Purpose of the Study:

• To develop a predictive model for forecasting student performance (grade and engagement) using online learning data.
• To analyze the impact of specific online learning platform features on student outcomes.
• To leverage machine learning for enhanced educational analytics.

Main Methods:

• Utilized a quantitative approach for analyzing student data from online learning platforms.
• Implemented machine learning techniques, specifically the Random Forest classifier, for predictive modeling.
• Incorporated student profile information and platform interaction data (e.g., discussion forums) as features.

Main Results:

• The Random Forest classifier demonstrated superior performance compared to other models.
• Achieved 85% accuracy in predicting student final grades.
• Achieved 83% accuracy in predicting student engagement levels.

Conclusions:

• Machine learning models can effectively predict student performance in online learning environments.
• Student profile and interaction data are significant predictors of academic success and engagement.
• The developed model offers a valuable tool for educators and institutions to monitor and enhance online learning.