Student Success Performance Prediction Using Tree-Based Pipeline Optimization Method (TPOM)

Predicting student success and analyzing student learning behaviors are critical components of improving online courses. By analyzing enormous amounts of clickstream data which tracks student behavior, instructors can get beneficial insights into the elements which impact students' academic results while recognizing opportunities for course development. In this research, a tree-based system is used in predicting student performance over academic for end-of-unit assignment through students' clickstream information. According to EDM Cup 2023 dataset, students with low capability are more probable to immediately ask for the response to an issue prior to intending a response, which has the potential to have detrimental effect on their learning results and might suggest attempts at manipulate the system. In the context of this purpose, we will anticipate growth in students while peering at the overall result through a device learning set of rules. The scholar's basic achievement over academic gets predicted using evaluating its earlier valuable data. When carrying out this prediction, the dataset utilized may assist in predicting student data. As a result, this study focuses on recognizing student success in academic performance utilizing the Tree-based Pipeline Optimization Method (TPOM) for Automatic Machine Learning (AutoML). In accordance with this pipeline, the Stochastic Gradient Descent (SGD) classifier performed better than other classifiers in the TPOM for AutoML in the EDM Cup 2023 dataset.