Predictive Analysis of Student Academic Performance in The Covid-19 Era: A Data Mining Approach

Abstract

Traditional educational methodologies were profoundly affected by the COVID-19 pandemic, which resulted in a complete shift towards remote and hybrid learning environments. This research seeks to analyze the effects of these new methodologies on student academic performance through predictive analysis and data mining techniques. By using a dataset that includes student population statistics, participation indicators, economic class, and academic performance, various machine learning techniques like decision trees, support vector machines, and neural networks are used to determine the most important predictors of student success. The relationship between the model's effectiveness in predicting academic performance, online learning adjustment, mental health, and digital accessibility is analyzed. The results indicate that ensemble learning achieves the best predictive accuracy, which supports the promise of a data-driven approach in educational policy making. The implications essentially serve as a new guide for educators and policy makers seeking remedies for adapting to post-pandemic student performance. This study highlights the value of incorporating predictive analytics in educational planning in order to reduce the adverse effects of interruptions in the learning system and assist students during a crisis.