An empirical study on effectiveness of temporal information as implicit ratings

Abstract

Collaborative filtering is a widely used and proven method of building recommender systems, which provide personalized recommendations on products or services based on explicit ratings from users. Recommendation accuracy becomes an especially important factor in some e-commerce environments (such as a mobile environment, due to limited connection time and device size). As user preferences change over time, temporal information can improve recommendation accuracy.This paper presents a variety of temporal information including item launch time, user buying time, the time difference between the two, as well as several combinations of these three. We conducted an empirical study on how temporal information affects the accuracy of a collaborative filtering system for recommending character images (wallpapers) in a mobile e-commerce environment. Empirical results show the degree of effectiveness of a variety of temporal information. The empirical results give insight on how to incorporate temporal information to maximize the effectiveness of collaborative filtering in various e-commerce environments.