|
|||||
|
|
| 拟合用户兴趣演变特性的协作过滤推荐算法 | |
| 引用本文： | 桑艳艳,刘培刚,李勇.拟合用户兴趣演变特性的协作过滤推荐算法[J].情报学报,2009,28(1). |
| 作者姓名： | 桑艳艳 刘培刚 李勇 |
| 作者单位： | 1. 南京大学信息管理系,南京,210093 2. 安徽财经大学商务学院,蚌埠,233030 3. 清华大学计算机科学与技术系,北京,100084 |
| 摘 要： | 个性化推荐技术是将传统的数据挖掘技术同用户访问信息结合起来,根据用户的兴趣爱好来对用户可能访问的内容进行预测并预取其提供给用户进行选择.目前协作过滤技术是个性化推荐系统中应用最为成功的推荐技术之一,但传统的协作过滤算法没有考虑用户的兴趣演变,难以有效地反映用户真实兴趣.在分析目前协作过滤算法存在问题的基础上,利用用户访问兴趣分为偶然兴趣和稳定兴趣的特性,文章提出了基于偶然兴趣的推荐权重和基于稳定兴趣的推荐权重,并将它们融入新的拟合用户兴趣演变的协作过滤算法中.实验表明该算法能准确地反映用户访问兴趣,较传统的协作过滤算法可以有效提高推荐精度. |
| 关 键 词： | 兴趣演变 协作过滤 个性化推荐 |
A Collaborative Filtering Algorithm Fitting User Interest Evolution |
|
| Sang Yanyan,Liu Peigang,Li Yong.A Collaborative Filtering Algorithm Fitting User Interest Evolution[J].Journal of the China Society for Scientific andTechnical Information,2009,28(1). | |
| Authors: | Sang Yanyan Liu Peigang Li Yong |
| Institution: | 1.Department of Information Management;Nanjing University;Nanjiag 210093;2.School of Business;Anhui University of Finance & Economics;Bengbu 233030;3.Department of Computer Science and Technology;Tsinghua Uviversity;Beijing 100084 |
| Abstract: | Personalized recommendation combines the data mining technology with user's browse profile and provides recommendation set to user forecasted by their interests.Now collaborative filtering is the most successful recommendation technology,but traditional collaborative filtering algorithms don't consider user's interest evolution and can't reflect user's true interests.On the analysis of the present algorithms for mining user's interests,two weights based on user's stable interests and accident interests are ... |
| Keywords: | interest evolution collaborative filtering personalized recommendation |
| 本文献已被 CNKI 万方数据 等数据库收录！ | |