E-Learning Outcomes of Engineering College Students Prediction Model Based on Machine Learning Technique

Abstract

This study employs examined machine learning techniques to predict e-learning outcomes for engineering college students. Specifically, it focuses on the application of Decision Tree and Random Forest models to predict student performance metrics such as dropout rates and final grades. Utilizing a comprehensive dataset compiled from various sources, including student demographics, academic histories, course interactions, and assessment scores, the study aims to identify patterns and key predictors of academic success in an online learning environment. The models are trained and evaluated using a robust dataset, with performance metrics such as accuracy, precision, recall, and F1-score serving as benchmarks for model effectiveness. The Decision Tree model provides an intuitive understanding of the data by illustrating the decision paths based on feature importance, while the Random Forest model enhances prediction accuracy through ensemble learning, effectively managing class imbalances and complex data structures. Findings from this study reveal significant insights into factors influencing student performance, offering potential strategies for educational interventions. The research highlights the capability of machine learning models to not only forecast educational outcomes with considerable accuracy but also to empower educational institutions with data-driven tools for enhancing student engagement and educational planning. The predictive models developed and tested in this study demonstrate a promising approach to addressing challenges in e-learning systems, ultimately aiming to improve the educational achievements of engineering students in an online setting.