Computer supported cooperative learning (CSCL) has attained considerable attention in recent years, but most CSCL systems do not consider ways of supporting learners in finding appropriate learning companions. In this study, we propose an intelligent learning companion recommendation mechanism (ILCRM) to deal with this problem. Specifically, ILCRM comprises three agents: (i) a candidate retrieval agent (CRA), (ii) a candidate evaluation agent (CEA), and (iii) a GA-based learning companion composition agent (GLCCA). The CRA and CEA are used to search a series of learning companion candidates based on two criteria (expertise level and participation level), and the GLCCA is employed to compose an appropriate cooperative group in which group members could be able to help learners solve the problems they face. The experimental results show that the proposed approach obtains a near optimal learning companion recommendation, has a significantly low computational cost, and satisfies the specified demands.
Multimedia Tools and Applications, v.76, n.9, pp.11547-11565