Predicting Student Performance from Online Engagement Activities Using Novel Statistical Features

Area of Science:

• Educational Technology
• Machine Learning
• Data Science

Background:

• The COVID-19 pandemic accelerated e-learning adoption, increasing online learning data availability.
• Predicting student performance is crucial for institutional decision-making and improving student outcomes.
• Machine learning models are increasingly used to analyze online learning data for performance prediction.

Purpose of the Study:

• To develop and evaluate a machine learning model for predicting student performance in online interactive sessions.
• To identify key features from student interaction data for accurate performance prediction.
• To compare the effectiveness of different machine learning classifiers for this prediction task.

Main Methods:

• Collected a dataset tracking student interactions during online lab work (e.g., text editing, keystrokes, time spent).
• Extracted 86 novel statistical features categorized into activity type, timing statistics, and peripheral activity count.
• Utilized feature selection to retain influential features and trained five classifiers: Random Forest, Support Vector Machine, Naïve Bayes, Logistic Regression, and Multilayer Perceptron.
• Evaluated model performance using random data split, cross-validation, and leave-one-session-out testing.

Main Results:

• The Random Forest classifier achieved the highest classification accuracy of 97.4%.
• The proposed model demonstrated superior performance compared to existing studies under similar experimental conditions.
• Feature selection identified influential factors for predicting student performance in online settings.

Conclusions:

• Machine learning models, particularly Random Forest, can effectively predict student performance in online learning environments.
• Detailed analysis of student interaction data provides valuable insights for educational interventions.
• The developed model offers a robust tool for institutions to monitor and enhance student success in e-learning.