Predictive Monitoring of Learning Processes

Abstract

What students do in a self-paced online learning environment is a "black box". The instructor has limited interactions with students and a restricted understanding of how students are progressing in their studies. A technology, sophisticated enough to predict the outcome of the student in an online learning environment was widely adopted in Predictive Learning Analytics. In the past, research on predictive learning analytics has emphasized predicting learning outcomes rather than facilitating instructors and students in decision-making or analyzing student behavior. This research study employed a predictive process monitoring technique to analyze the student’s event logs in an online learning and online test environment to predict the next activity the student is going to perform and the remaining time to complete the course or test. The Long Short Term Memory neural network approach is used in this work to predict the next activity of the running case by analyzing the sequence of historical data and Apromore to predict the completion time of a case. By employing the predictive monitoring of learning processes, new insights are developed to analyze students’ behavior in real-time and is achievable.