Skip to Main Content
Recommender systems provide the user with the items that they might wish to purchase. Most current recommendation systems are too dependent on the users. The users are often required to rate different items and even give comments on the products they purchased. Worst of all, when the user's preferences are changed, the whole recommender systems are usually not adaptive to the new condition. In this paper, an intelligent recommender system architecture is proposed. In this architecture, an autonomous agent is used to track the user's Web browsing behaviours and attempt to anticipate the user's current interests to build up the user profile in a dynamic way. On the other hand, an explicit item model has been applied to each item and item-to-item correlations have been used to assist the agent to give better recommendations.