A Collaborative Filtering Method Based on the Forgetting Curve

Collaborative filtering (CF) is one of the most successful approaches for making personalized recommendations. This paper aim sat the issue of tracking the drifting of the user´s interests, and proposes a novel collaborative filtering recommendation method based on Ebbinghaus Forgetting Curve. The new method learns and tracks the user´s interests by defining the user´s interests as the short-term interest and the long-term interest, and by defining the weight function based on the time-window as well. In order to produce high quality recommendations, both the data weight based on the time-window and the data weight based on the item-similarity are used. Furthermore, the paper finds a special power function curve is much more fit to the forgetting curve through a mathematical analysis tool. Comparative experiments with the standard data show that the proposed method providing dramatically better quality recommendations.