Awareness based filtering - Toward the cooperative learning in human agent interaction -

We propose a new method of the recommender agent that supports to make clear the preference of a user. First, an overview of our coarse to fine recommendation system and the experimental results are described. Next we will discuss the effectiveness of our system. We interpret the meaning of our fixed recommendation strategy of the agent for contributing to restrain to prevent the user from his awareness. Then we will propose the awareness based filtering method to assist a user to be aware the true preference of the user. The main idea is to visualize both the coarse to fine preference space and the history of the recommendation in it.