Predicting Student Academic Performance Using Feature Engineering on E-Learning Platforms

Abstract

The widespread adoption of e-learning platform has transformed modern education by enabling continuous monitoring of students engagement, learning behavior, and academic performance. Learning management system [LMS]. Such as moodle, Coursera and Edx collect large volume of behavioral data including login, activity, resource interaction, assignment, submission and discussion. Forum Participation. These datasets provide valuable insights that can be analyzed using machine learning algorithms to predict student academic outcomes and identify learners at risk of academic failure. However, raw LMS interaction data is often noisy, inconsistent and difficult to interrupt, which limits the Reliability of predictive models Feature Engineering plays a critical role in transforming raw behavioral logs into meaningful indicators such as study consistency, time on tasks, participation intensity, and learning persistence Students using real world data. Let's demonstrate that Engineered features. Significantly improved predicting, accuracy and interpretability of machine learning Model. This research analyzes how feature engineering enhances academic performance predicting models while maintaining transparency. Fairness and ethical AI adoption in education. The study synthesizes binding from recent research to propose a conceptual framework that supports Interpretable predictive analytics aligned with responsible AI principles in educational environment.