Group Prediction in Collaborative Learning

Abstract

We propose an approach for predicting group formations, to address the problem of automating the incorporation of group awareness into CSCL applications. Contextual information can enable the construction of applications that effectively assist the group members to automatically communicate in synchronous and collocated collaborative learning activities. We used data traces collected from the study of students’ behavior to train and test an intelligent system. Results have shown that context-information can be effectively used as a basis for a middleware for a dynamic group management. Inferring group membership is technically viable and can be used in real world settings.