While making significant advances in the recommender system technology, most of the prior research studies in this topic area have focused on providing recommendations for individuals, which unfortunately can not be effectively applied for group recommendations, i.e., making recommendations for a group of people. Notice that a common idea behind recommender systems that make personalized recommendations for individuals is to discover users' preference profiles (either from user ratings or item text descriptions) in order to identify items that best match the profiles of targeted users. However, for group recommendations, we argue that a good recommender system not only needs to model users' individual preferences but also understands how a decision among group members is reached.