A Collaborative Filtering Recommendation Algorithm Based on Interest Forgetting Curve

Abstract Collaborative filtering (CF) algorithm is one of the most successful technologies used in personalized recommendation system. However, traditional algorithms focus only on the user ratings and do not take the changes of user interest into account, which affect recommendation quality seriously. To address the issue, this paper proposes a CF algorithm based on interest forgetting curve. Based on a specially designed experiment, it first explores the law of user interest changing—interest forgetting curve. Then, it uses recently rated items to represent the user’s current interest; for each historically visited item, it calculates the integrated data weight based on interest forgetting curve and the user-item rating matrix; for each item without the user’s score, it calculates prediction based on item similarity and item integrated data weights. While calculating items’ similarity, it combines the item attribute similarity and item score similarity, which is more comprehensive and accurate. The experimental results show that the proposed algorithm can provide better recommendation precision and recall ratio.