Abstract
Virtual classrooms have become an essential part of today's educational scene because of their accessibility and flexibility. For instructors working in these settings, guaranteeing student engagement and academic performance is still a major concern. Using predictive analysis tools, this project seeks to improve the virtual classroom experience by gaining information about student performance and engagement. Through behavior prediction, teachers can modify their approach to teaching and create a more engaged and productive learning environment. This research uses machine learning algorithms like logistic regression and random forest classifiers, using datasets from an undergraduate science course in the second year at North American University. These algorithms are used to predict engagement levels based on activity measurements and student success metrics including course grades, pass/fail results, attendance, and individual exam scores.
With the use of predictive models, teachers can better understand student behavior and academic performance and adapt their teaching strategies to better match the needs of the students. Teachers can effectively see and evaluate student data in an interactive dashboard by integrating these predictions into a front-end Flask application. This research offers a unique method for enhancing the virtual classroom experience using predictive analysis. Through the utilization of machine learning, instructors can acquire a significant understanding of student conduct and involvement, which can ultimately result in more customized and efficient instructional strategies.