An Improved E-learner Community Construction Algorithm Based on Learning Interest Feature Vectors

Finding similar e-learners in a distributed and open elearning environment and help them to learn collaboratively is becoming one of the urgent challenges of personalized e-learning services. Literature shows that current e-learner community building approaches are generated from qualitative studies of small-sized learner-centered classrooms which may need the teacher's participation. However, the findings might not apply to large classes in distributed learning environments, which make the teachers to face hundreds of e-learners in each class. In such situations, teachers also find it impossible to analyze the learning behaviors of each e-learner and divide them into different learning communities accurately. This paper addresses this problem in the adaptive e-learner community self-organizing point of view. Considering both the feature vector of learning resources and a learner's rating value on each resource, this paper firstly defines the learning interest feature vector to model the learner's behavior. Based on this an accurate learning interest feature representation method and, an innovative e-learner community self-organizing algorithm, called IFVSORC, are proposed in this paper. Experimental results show that this algorithm exhibits good community organizing efficiency and scalability. Key-Words: E-Learning, Collaborative Learning, Community Construction, Interest Feature Vector