南京大学学报(自然科学版) ›› 2013, Vol. 49 ›› Issue (2): 142149.
韦素云1**,业宁1,吉根林2,张丹丹1,殷晓飞1
Wei Su-Yun1 ,Ye Ning1,Ji Gen一Lin2,Zhang Dan Dan1 ,Yin Xiao Fei1
摘要: 用户评分数据极端稀疏情况下,传统相似性度量方法存在弊端,导致推荐系统的推荐质量急
剧卜降.针对上述问题,提出一种基于项目类别和兴趣度的协同过滤推荐算法.在该算法中,首先通过计
算项目之间的类别距离,构造项目类别相似性矩阵;然后采用兴趣度分析不同项目之间的相关程度;最
后结合项目类别信息和项目间的兴趣度,使用改进的条件概率方法作为衡量项目间相似性的标准.实验
结果表明,该算法可以有效缓解用户评分数据稀疏带来的不良影响,提高预测准确率和推荐质量.
[1]Breese J,Hecherman D,Kadie C. Empirical analy- sis of predictive algorithms for collaborative filte- ring. Proceedings of the 14th Conference on Un- certainty in Artificial intelligence. San Francisco: Morgan Kaufmann Publishcrs,1998,43一52. [2]Linden G, Smith B, York J. Amazon, com recom mendations; Item to item collaborative filtering. IEEE lnternet Computing,2003,7(1):76一80. [3]Das A,Datar M,Garg A,et al. Google news per- sonalization; Scalable online collaborative filte- ring. Proceedings of the 16th international Confet- ence on World Wide Web. New York; ACM,2007,271一280. [4]George T,Merugu S. A scalable collaborative fil tering framework based on co-clustering. Pro ceedings of the 5th Conference on Data Mining. Washington; IEEE Computer Society Press 2005,625一628. [5]Vozalis M, Margaritis K. Applying SVD on item- based filtering. Proceedings of the 5th Conference on intelligent Systems Design and Applications,Washington; IEEE Computer Society Press,2005,464一469. [6]Zhang S, Wang W H,Ford J,etal. Learning from incomplete ratings using non-negative matrix far torization. Proceedings of the 6th Conference on Data Mining. Bethesda;SlAM,2006,549一553. [7]Aggarwal C C, Wolf J L,Wu K L, et al. Homing hatches an egg; A new graplrthcoretic approach to collaborative filtering. Proceedings of the 5th international Conference on Knowledge Discovery and Data Mining, New York; ACM Press, 1999,201-212. [8]Adomavicius G,Tuzhilin A.Toward the next generation of recommender systems; A survey of the state-of-thwart and possible extensions, IEEE Transactions Knowledge and Data Engineering, 2005,17(6):734一749. [9]Zeng C, Xing C X,Zhou L Z, et al. Similarity measure and instance selection for collaborative filtering.International .Iournal of Electronic Com- merce,2004 ,8(4):115一129. [10]Zhou J F,Zang X,Guo J F. An optimized collaborative filtering recommendation algorithm. Journal of Corm puter Research and Development, 2004,41(10):1842 -1847.(周军锋,汤显,郭景峰.一种优化的协 同过滤推荐算法.计算机研究与发展,2004,41 (10):1842一 1847. [11]Li C, Liang C Y, Dung K. A collaborative filtering recommendation algorithm based on item category similarity. Journal of Hefci University of Zcchnology ( Natural Sciences),2008,31(3):360一363.(李 聪,梁昌勇,董坷.基于项目类别相似性的协同过滤推荐算法.合肥工业大学学报(自然科学),2008,31(03):360一363. [12]Yu H,Li Z Y. A collaborative filtering recom- mendation algorithm based on forgetting curve. Journal of Nanjing University(Natural Sciences),2010,6(5):520-527.(于洪,李转运.基于遗 忘曲线的协同过滤推荐算法.南京大学学报(自然科学),2010,46(5):520 -527). [13]Sarwar B M, Karypis G, Konstan J,et al. Item- based collaborative filtering recommendation al- gorithms. Proceedings of the 10th Conference on World Wide Web. New York:ACM Press,2001,285一295. [14]Sarwar B M. Sparsity,scalability,and distribution in recommender systems. Minneapolis; University of Minnesota,2001. [15]Piatesky S G Discovery, analysis and presentation of strong rules. Knowledge Discovery in Database. Menlo Park , MlT,1991,229一248. [16]Karypis G. Evaluation of item-bascd top-n recom- mendation algorithms. Proceedings of the 10th Con- ference on information and Knowledge Manages ment. New York:ACM Press,2001 ,247一254. [17]Deshpande M, Karypis G. item-based top-n rec-ommendation algorithms. ACM Transactions on Information Systems,2004,22( 1):143一177. |
No related articles found! |
|