Adapting recommendations organization to user preferences

Given the overwhelming information appearing in the current web environment, recommendations have been increasingly applied to assist users in handling with the information overload and locating items that interest them. As a different way to generate and display recommendations, the organization interface has been found being more effective in building users' trust. In this paper, we propose a novel approach to generating the organization of recommendations, with the goal of making it dynamically adaptive to different conditions of user preferences (i.e. "incomplete preferences" and/or "preference conflicts"), so as to optimally support users' preference construction process and accurate decisions.