The paper introduces a novel feature selection technique, BR^2-2 T, which combines Ridge (L2) regularization and Boruta optimization to enhance prediction accuracy in e-learning. This two-tier approach is integrated with a three-tier ensemble model that includes Random Forest, Support Vector Machine (SVM), and Gradient Boosting, each optimized using Bayesian Optimization, random search, and Particle Swarm Optimization (PSO) respectively. The study also employs Z-score normalization, Synthetic Minority Over-sampling Technique (SMOTE), and Multiple Imputation by Chained Equations (MICE) to handle data issues. The proposed method achieves a maximum accuracy of 98.74%, significantly outperforming traditional methods in educational outcome prediction. The research underscores the importance of advanced algorithms in improving student success and educational pedagogy. The introduction highlights the benefits of e-learning, including flexibility, accessibility, and adaptability, and discusses the growing global market and student usage of online learning platforms. The literature survey reviews recent studies on machine learning models for specialized choice, learning preference detection, and virtual learning environment behavior analysis, emphasizing the potential of these techniques in enhancing educational outcomes.The paper introduces a novel feature selection technique, BR^2-2 T, which combines Ridge (L2) regularization and Boruta optimization to enhance prediction accuracy in e-learning. This two-tier approach is integrated with a three-tier ensemble model that includes Random Forest, Support Vector Machine (SVM), and Gradient Boosting, each optimized using Bayesian Optimization, random search, and Particle Swarm Optimization (PSO) respectively. The study also employs Z-score normalization, Synthetic Minority Over-sampling Technique (SMOTE), and Multiple Imputation by Chained Equations (MICE) to handle data issues. The proposed method achieves a maximum accuracy of 98.74%, significantly outperforming traditional methods in educational outcome prediction. The research underscores the importance of advanced algorithms in improving student success and educational pedagogy. The introduction highlights the benefits of e-learning, including flexibility, accessibility, and adaptability, and discusses the growing global market and student usage of online learning platforms. The literature survey reviews recent studies on machine learning models for specialized choice, learning preference detection, and virtual learning environment behavior analysis, emphasizing the potential of these techniques in enhancing educational outcomes.