收稿日期: 2016-10-08
修回日期: 2016-12-09
网络出版日期: 2016-12-20
基金资助
本文系校科研计划项目“基于机器学习的评论探测应用研究”（项目编号：GYKY/2016/15）研究成果之一。
Application of an Improved Collaborative Filtering Method on Recommending Books in College Libraries
Received date: 2016-10-08
Revised date: 2016-12-09
Online published: 2016-12-20
[目的/意义] 为解决高校图书推荐过程中面临的“数据稀疏”和“冷启动”问题，研究表明：优化读者评价矩阵和相似度模型是提高图书推荐质量的关键。[方法/过程] 提出一种协同过滤改进方法，以图书分类为项目生成用户评价矩阵，并考虑借阅方式、借阅时间和图书相似度对用户兴趣度的影响，优化矩阵中的样本数据；同时，在计算读者相似度时融入读者特征和图书特征。[结果/结论] 实验结果表明，该方法可有效解决“数据稀疏”和“冷启动”问题，显著降低计算量。与基本协同过滤和聚类改进的协同过滤方法相比，无论是在推荐准确率还是在用户满意率上都有较大的提高，综合推荐效果更好。
宋楚平 . 一种改进的协同过滤方法在高校图书馆图书推荐中的应用[J]. 图书情报工作, 2016 , 60(24) : 86 -91 . DOI: 10.13266/j.issn.0252-3116.2016.24.012
[Purpose/significance] In view of the problems of "sparse data" and "cold start" in the process of book recommendation in the colleges or universities, optimizing the readers' evaluation matrix and the similarity model is the key to improve the quality of book recommendation.[Method/process] An improved collaborative filtering method is proposed based on the classification of books to generate user evaluation matrix, and optimizes sample data in the matrix by considering the influence of the lending type, time and the similarities of the book on the user's interest. In addition, characteristics of readers and books are taken into consideration in calculating the reader's similarity, which can effectively solve the the problems of "sparse data" and "cold start" and significantly reduce the amount of computation.[Result/conclusion] The experimental results show that the improved method is better than the basic collaborative filtering and clustering collaborative filtering method, and it has a great improvement on the recommendation accuracy and user satisfaction, thus the effect of integrated recommendation is better.
[1] 李默,梁永全.基于标签和关联规则挖掘的图书组合推荐系统模型研究[J].计算机应用研究,2014,31(8):2391-2392.
[2] 马仲兵.基于关联规则的高校图书馆个性化推荐模型[J].新世纪图书馆,2013(7):42-43.
[3] 李欣.基于概念格的图书关联规则挖掘研究[J].图书馆学研究,2015(8):73-74.
[4] YANG B,NIGEL C,ANINDYA D.A partially supervised cross-collection topic model for cross-domain text[C]//Proceedings of the 22nd ACM international conference of information & knowledge management(CIKM 2013).New York,USA:ACM,[S1]:[s.n.],2013:240-246.
[5] 郑祥云,陈志刚,黄瑞,等.基于主题模型的个性化图书推荐算法[J].计算机应用研究,2015,35(9):2570-2573.
[6] 马甲林,张永军,王志坚.基于概念簇的多主题提取算法[J].智能系统学报,2015,10(2):261-264.
[7] 孙彦超.基于聚类分析算法的图书推荐系统的研究[J].图书馆理论与实践,2015(5):76-78.
[8] 张闪闪,黄鹏.高校图书馆图书推荐系统中的稀疏性问题实证探析[J].大学图书馆学报,2014(6):47-50.
[9] 王连喜.一种面向高校图书馆的个性化图书推荐系统[J].现代情报,2015,35(12):41-44.
[10] 郝立燕,王靖.基于填充和相似性信任因子的协同过滤推荐算法[J].计算机应用,2013.33(3):834-837.
[11] 张玉芳,代金龙,熊忠阳.分步填充缓解数据稀疏性的协同过滤算法[J].计算机应用研究,2013. 30(9):2602-2605.
[12] GARCIA F,ROMCRO C,VENTUTA S, et al. An architecture for making recommendations to courseware authors using association rule mining and collaborative filtering[J].User Modeling and User-Adapted interaction,2009,19(1-2):99-130.
[13] SARWAR B, KARYPIS U, KONSTAN J,et al. Analysis of recommendation algorithms for E-commerce[C]//Proceedings of the ACM E-Commerce. NewYork,USA:ACM Press, 2000:158-162.
[14] LEUNG C W K,CHAN S C F, CHUNG F L.An empirical study of a cross-level association rule mining approach to cold-start recommendations[J].Knowledge-based systems,2008,21(7):515-527.
[15] SHAW G, XU Y,GEVA S. Using association rules to solve the cold-start problem in recommender systems[C]//Proceeding of the 14th Pacific-Asia Conference On Advances in Knowledge Discovery and Data Mining.Berlin:Springer, 2010:340-347.
[16] BALTRUNAS L,AMATRIAIN X.Towards time-dependent recommendation based on implicit feedback[C]//Workshop on context-aware recommender systems (CARS'09).New York:ACM,2009.
[17] GOLDBERG D,NICHOLS D,OKIB M,et al.Using collaborative filtering to weave an information tapestry[J].Communications of the ACM,1992,35(12):61-63.
/
| 〈 |
|
〉 |
