Student GroupRecommendation using Extreme Learning Machinefor Collaborative Learning

Abstract

Collaborative learning(CL) is defined as situations where several learners with different ways of thinking, acting, and feeling participate in solving common problems. Group work is a representative form of CL and a strong need for various online courses. In this study, we focus on the student group recommendation based on the personalized information of students, including learning styles and social closeness. Moreover, aneffective recommendation method using extreme learning machine (ELM) is proposed in this paper. The method regards studentgroup recommendation asa binary classification problem. First, learning features are extracted from group learning styles using Felder-Silverman model and group members' social relationships. These features simultaneouslyconsider all the important factors which determine the results of recommendation and guarantee the effectiveness of grouprecommendation. Then, the extracted features are input to train ELM classifier due to its fast learning speed, whichguarantees the efficiency of recommendation. Finally, comparative experiments verify the accuracy and efficiencyof our method.