Learning Analytics for Predicting Academic Performance in Computer Science Courses

Abstract

The acclaimed development of digital learning environment has produced massive volumes of pedagogical information, which has provided the chances to enhance the outcomes of the students using learning analytics. The paper explores the application of learning analytics in how academic performance is forecasted in undergraduate students pursuing Computer Science courses. The predictive insights were created by analyzing the variables of attendance, learning management system (LMS) participation, patterns of assignments submission, quiz grades, and programming lab activity, to detect at-risk learners. Data obtained by 200 undergraduate students was subjected to a quantitative research method with correlation and regression to get the relationship between smoking and ADHD. The results indicate that there is a great positive association between LMS engagement indicators and final academic performance. The models of early prediction proved to be very accurate in determining the students who were likely to perform poorly in order to initiate the academic intervention early. The paper highlights the necessity of incorporating the use of data-based decision-making tools into the Computer Science education system to improve student performance, decrease the dropout rates, and promote individualized learning experiences.