Enhancing Student Performance Prediction with Greylag Goose Optimization Algorithm

The current paper addresses the central role of hyperparameter optimization in improving the predictive power of the MLP Regressor for forecasting student performance in Portuguese secondary schools. The uniqueness of this research lies in its exploration of metaheuristic optimization algorithms, specifically highlighting GGO (Greylag Goose Optimization) for enhancement. The study utilized a dataset crucial for understanding and predicting student performance, with a special focus on its distinctive features. By comprehensively tuning the MLP Regressor, the paper demonstrates remarkable improvements in various performance measures, as evident in the enclosed tables. Specifically, the MSE values calculated for the MLP Regressor both before and after GGO optimization are compared. Without optimization, the MLP Regressor had an MSE of 0.0103. After GGO optimization, the MSE significantly improved to 0.0060, indicating enhanced accuracy with GGO in the model. These findings emphasize that hyperparameter optimization, particularly the GGO technique, is crucial for refining the MLP Regressor in predicting student performance. The paper not only delves into the technical aspects but also concludes by highlighting the broader consequences of these outcomes. The potential educational applications are substantial, as improved models can provide more accurate predictions, empowering educators and policymakers to make informed decisions in education. This paper establishes a foundation for future research directions, contributing to the pool of ideas for educational predictive modeling.