A Time Effect based Collaborative Filtering Approach for User Preference Statistics and Recommendation

Yuxi Chen, Xiaotong Zhang, Qing Zhao*, Faheem Akhtar

*Corresponding author for this work

Research output: Contribution to JournalArticleAcademicpeer-review


With the rapid development of the information technologies in the financial field, extracting meaningful information from a massive amount of data is hugely significant for efficient business decision making. The recommendation system is an intelligent system that applies historical knowledge of users to infer their preferences and make a personalized recommendation. However, it suffers from the problem of time effect of user's behaviour, which means a user's interests may change over time. To overcome this problem, we propose a time effect based collaborative filtering approach to adaptively statistics the change of user preferences. Firstly, Item-based collaborative filtering is used to calculate rating similarity between items. Since an Item-based collaborative filtering algorithm doesn't consider the time effect; next, the time decay function is proposed to statistics the change of user interests. Experimental results show that the proposed scheme retained higher accuracy compare to traditional collaborative filtering method.

Original languageEnglish
Pages (from-to)2-7
Number of pages6
JournalJournal of Physics : Conference Series
Publication statusPublished - 3 Mar 2020
Event2019 2nd International Conference on Computer Information Science and Artificial Intelligence, CISAI 2019 - Xi'an, China
Duration: 25 Oct 201927 Oct 2019


Dive into the research topics of 'A Time Effect based Collaborative Filtering Approach for User Preference Statistics and Recommendation'. Together they form a unique fingerprint.

Cite this