Abstract
In contemporary MOOC (Massive Open Online Courses), videos take a central role in delivering course content. The analysis of how students interact with these videos becomes critical to understand how student learn in such environments. We analyze sequences of video interactions of 4800 students with 79 videos corresponding to the half time course (six weeks) of an online course of thirteen weeks to predict student performance at the end of the course. We use an adapted modified transition matrix representation of student video interactions (TMED) and use four classifiers such as Support Vector Machine (SVM), Gradient Boosted Machine (GBM), K-Nearest Neighbors (KNN) and Random Forest (RF) to analyze student video interactions. To validate the proposed method, we first test the capacity of TMED to discriminate individual student video interactions and, second, verify if one can predict student performance at the end of the course based only on the first week of student video interaction and then considering the interaction with videos up to the mid-course. The results show that this method based on TMED perform better than two previous studies based on video utilization measures of student and rules of classification.
Recommended Citation
Mbouzao, Boniface, "Student Performance Prediction in MOOC using Students' Video Interaction Analysis." (2024). CACAIS 2024 Proceedings. 11.
https://aisel.aisnet.org/cacais2024/11