An Analysis of Learning Behavior Patterns with Different Devices and Weights

Abstract

With e-learning systems gradually being implemented, researchers worldwide have started devoting increasing attention to Learning Analytics. At Kobe university, a digital textbook reading system has been developed to collect learning logs in the face-to-face classroom. In a previous study, k-means clustering was implemented to analyze learning behavior patterns; however, there were problems such as few variables for clustering and a failure to consider weighting of the learning elements. Therefore, in this study we applied clustering by increasing the number of learning elements and assigned weights to the learning elements, then analyzed the learning behavioral patterns. We found some behavioral patterns of students who can save learning time if they effectively write memos and add markers.