摘要
传统推荐算法没有考虑时间效应的影响,而随着用户兴趣、产品流行度等变化,会使得推荐效果受到影响。近年来,越来越多的研究者开始关注推荐系统动态特性,时间信息对推荐系统有重要的作用,将回顾推荐系统主要算法,研究静态模型存在的问题,详细介绍近年来国内外动态推荐算法的研究进展,为后续研究提供参考。
Earlier recommender system did not take temporal effects into account.But accompanied by the change of user interests and item popularity,the performance of recommender system isn't so good.Therefore,in recent years more and more researchers realized that temporal information play an extremely important role in recommender systems.This paper reviews the methods of recommender system,and discusses what's the main problem of static model without time information.Furthermore,introduces the progress of time-based dynamic recommender system research in recent years.
出处
《工业控制计算机》
2015年第8期99-100,103,共3页
Industrial Control Computer
关键词
推荐系统
协同过滤
时间效应
兴趣变化
动态特性
recommender system
collaborative filtering
temporal effects
interests changing
dynamic features