摘要
下一个购物篮推荐旨在为用户推荐下一次想购买的商品,与传统推荐不同,它考虑了商品间的时序相关性,购物篮中的商品可能相互依赖,而非彼此独立。如何更好地挖掘用户的时序行为特征是下一个购物篮推荐的难点和挑战。该文首先阐述了下一个购物篮推荐的内涵及研究意义,并对它做基本介绍。然后重点梳理了下一个购物篮推荐的研究进展,对其算法整理分析,阐述每种方法的优势和不足,并贯穿了其算法的发展情况、存在问题和研究现状。最后对下一个购物篮推荐的发展进行展望,为后续研究提供参考。
Next basket recommendation is designed to recommend a basket of items that the user most probably would buy next time.The main difference between it and the traditional recommendation is that sequential dependencies among items are con⁃sidered,items in basket may be dependent rather than independent.How to elicit user's sequential behavior features better is the dif⁃ficulty and challenge of next basket recommendation.This paper first describes the meaning and significance of next basket recom⁃mendation,introduces next basket recommendation.Then it focuses on the research progress of next basket recommendation,ana⁃lyzes its algorithms,illustrates the pros and cons of each method,and runs through the development,existing problems and re⁃search status of its algorithms.Finally it makes the prospect for the development of next basket recommendation,provides reference for future study.
作者
王伟玉
WANG Weiyu(Institute of Computing Technology,Chinese Academy of Sciences,Beijing 100190;University of Chinese Academy of Sciences,Beijing 100190)
出处
《计算机与数字工程》
2021年第2期340-346,共7页
Computer & Digital Engineering
关键词
下一个购物篮推荐
时序相关性
个性化推荐
next basket recommendation
sequential dependency
personalized recommendation