Collaborative filtering algorithm with social information and dynamic time windows

With the rapid development of social networks, the problem of information overload is increasingly serious. The recommendation system can deal with the problem of information overload effectively and provide users with personalized recommendation services. In the process of recommendation, the traditional recommendation algorithms do not take the social relationship of users as the basis of recommendation; at the same time, they do not take for the dynamic change of user's interest and think that it is immutable. About these problems, the paper proposes a personalized recommendation algorithm with social information and dynamic time windows. Firstly, a collaborative filtering algorithm is proposed which integrates social information and user interest in the process of searching the nearest neighbor. Secondly, the time windows are dynamically adjusted to obtain a stable increment and better reflect the short-term interests of users. Then, the concept of time function is introduced to allocate corresponding time weights for users' interests in different periods. Finally, we conduct a series of experiments to verify the practicability and effectiveness of our algorithm. Experimental results show that the performance of the proposed algorithm is better than the traditional collaborative filtering recommendation algorithm.

Automated summary

This paper proposes a personalized recommendation algorithm that integrates social information and dynamic time windows. The algorithm adjusts time windows dynamically to reflect users' short-term interests, and uses social information in the process of searching the nearest neighbor to achieve better performance compared to traditional collaborative filtering recommendation algorithms.