Abstract: Recommender systems are used to help users find relevant and personalized items from a large set of alternatives in many online applications．Most existing recommendation techniques are focused on improving recommendation accuracy．Recently,diversity of recommendations has also been increasingly recognized in research literature as an important aspect of recommendation quality．This paper proposed a novel top-N recommendations generated approach to improve aggregate recommendation diversity by controlling the recommended expectations of the global candidate items,which is available for recommendation in the top-N recommendations generating process．The proposed approach was evaluated using real-world movie rating datasets and different rating prediction algorithms．Results demonstrate the approach proposed in this paper can generate more diverse recommendations while maintaining an acceptable level of accuracy.

Key words: Recommender systems,Top-N recommendations,Diversity,Recommendation expectation