Analyzing Students Logs in Open Online Courses Using SNA Techniques

Leyla Zhuhadar, Jeff Butterfield

Journal of the Association for Information Systems · 2014 · 14 citations · 3 references

Concepts

Abstract

Abstract There are a growing number of courses delivered using Massive Open Online Course environments (MOOCs). In these courses, as with any, interaction between online students and learning resources can play in improving the learning environment. These improvements could better accommodate each individual’s needs based on their prior knowledge and skills. Most current MOOCs use a one-size-fits-all model with a singular curriculum. We suggest that a singular curriculum does not provide a particularly good match for this broad spectrum of students with diverse educational backgrounds. This problem is manifest in the relatively low rates of completion in many MOOC courses. As an alternative, we propose the notion of Personalized Open Collaborative Courses (POCCs). This design tracks student behavior through each stage of the course and, based on each student’s performance and prior-knowledge, the course content and delivery is personalized. This study follows the emerging research in Social Network Analysis (SNA) techniques. More specifically, we examine the concept of detecting communities of users within a large course with an emphasis on designing a personalized, social, recommender system. We use three of the most popular community detecting algorithms: Fruchterman Reingold, ForceAtlas2 and Yi Fan Hu for community detection. Finally, we define an algorithm to build dynamic social recommender systems.

References

3