Group recommendation system for e-learning communities: A multi-agent approach

Abstract

For many web services, recommendation system plays a very important role. In most situations, those recommendations are personalized to individual users. However in some services, like e-learning services, not all activities may be designed for personal usage, but rather to be conducted in a group. Group recommendation has been a very challenging problem to address due to the different preferences amongst members of a group. In this work, we propose a way to recommend courses to a group of e-learning students that considers the various preferences between students who happen to be online at the same time frame. We implement a multi-agent model to simulate the behaviors of those students and generate recommendations accordingly. By splitting the machine learning process across multiple intelligent agents, we found that our method managed to generate accurate recommendations.