摘要

Taking into account the characteristics of the recommendation of mobile services, this paper proposes a new collaborative filtering algorithm via clustering and applies the method in the recommendation of mobile services. By using the similarity of the users' past rating and the similarities of the users' profiles, the method is aimed to tow fundamental problems: accuracy and response speed. Empirical evaluation of the method on the purchase history of mobile services demonstrates that the method can enhance the real-time response speed as possible as without reducing the accuracy for the recommendation of mobile services.

全文