Mining online learner profile through learning behavior analysis

Abstract

User profile is an effective model to describe the user's interests and preferences. In the learning field, learner profile should meet the demand of learning and teaching such as learning patterns recognition or performance prediction. Analysis of user's behaviors is the common way to build user profile. Statistics show that online learning activities are incontinuous and diverse. By taking a closer look at the learning activity data, we found back accessing behavior is a frequent activity and reveals the truth of learners' intention. In this study, we make use of Shanghai Open University's learning platform as the data source for our research, adopt machine learning method to find the hidden patterns of learning activities and build the online learner profile. Statistics show that 15.68% of the accessing activities are back accessing. We found three learning patterns with different amount of back accessing behaviors and learning paths. Meanwhile, they relate to many factors including demographics, major type and area where learners join in learning. Through learner profile, we can predict learner's learning pattern which we found in this study. In the conclusion of our study, we suggest that learning path should be taken into consideration of learning engagement.