全网兴趣图谱下电子商务推荐理论研究现状  

An Exploration of E- commerce Recommendation Based on Interest Graph on Whole Internet

作　　者：陈媛媛[1,2] 聂规划[1,2] 陈冬林[1,2] 桂雁军[3] 

机构地区：[1]武汉理工大学经济学院,湖北武汉430070 [2]武汉理工大学电子商务与智能服务研究中心,湖北武汉430070 [3]新奇特车业服务股份有限公司,上海201824

出　　处：《武汉理工大学学报（信息与管理工程版）》2014年第1期109-113,共5页Journal of Wuhan University of Technology：Information & Management Engineering

基　　金：国家自然科学基金资助项目(71172043;71072077);国家科技支撑计划基金资助项目(2013BAH13F01;2012BAH93F04);教育部人文社科基金资助项目(09YJA630124;10YJAZH050);中央高校基本科研业务费专项资金资助项目(2013-ZY-131)

摘　　要：介绍了现有的电子商务推荐技术与系统,分析了其中存在的数据稀疏性、冷启动、推荐精度低和实时性差等问题,阐述了利用全网数据构建用户兴趣图谱进行推荐的方法,探讨了基于兴趣图谱的推荐与跨网推荐取得的进展,证明该方法的提出具有重要的现实意义。The existing e -commerce recommendation technology and systems were discussed. Some problems such as data sparsity, cold start, low precision and poor real - time capability in systems were analyzed. The method that constructing user in- terest graph was presented based on the data of whole internet. The progress of recommendation based on interest graph and cross - system was explored. It demonstrated that the proposition of the method has great practical significance.

关 键 词：全网 兴趣图谱 电子商务推荐 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]