资源描述：
《基于影响集的协作过滤推荐算法》由会员上传分享，免费在线阅读，更多相关内容在工程资料-天天文库。
1、ISSN1000-9825,CODENRUXUEWE-mail:jos@iscas.ac.cnJournalofSoftware,Vol.18,No.7,July2007,pp.1685−1694http://www.jos.org.cnDOI:10.1360/jos181685Tel/Fax:+86-10-62562563©2007byJournalofSoftware.Allrightsreserved.∗基于影响集的协作过滤推荐算法1+2陈健,印鉴1(华南理工大学计算机科学与工程学院,广东广州510006)2(中山大学计算科学系,广
2、东广州510275)ACollaborativeFilteringRecommendationAlgorithmBasedonInfluenceSets1+2CHENJian,YINJian1(SchoolofComputerScienceandEngineering,SouthChinaUniversityofTechnology,Guangzhou510006,China)2(DepartmentofComputerScience,SunYat-SetUniversity,Guangzhou510275,China)+Correspo
3、ndingauthor:Phn:+86-20-33509119,Fax:+86-20-39380218,E-mail:ellachen@scut.edu.cn,http://www.scut.edu.cnChenJ,YinJ.Acollaborativefilteringrecommendationalgorithmbasedoninfluencesets.JournalofSoftware,2007,18(7):1685−1694.http://www.jos.org.cn/1000-9825/18/1685.htmAbstract:T
4、hetraditionaluser-basedcollaborativefiltering(CF)algorithmsoftensufferfromtwoimportantproblems:Scalabilityandsparsitybecauseofitsmemory-basedknearestneighborqueryalgorithm.Item-BasedCFalgorithmshavebeendesignedtodealwiththescalabilityproblemsassociatedwithuser-basedCFappr
5、oacheswithoutsacrificingrecommendationorpredictionaccuracy.However,item-basedCFalgorithmsstillsufferfromthedatasparsityproblems.ThispaperpresentsaCFrecommendationalgorithm,namedCFBIS(collaborativefilteringbasedoninfluencesets),whichisbasedontheconceptofinfluencesetandisah
6、ottopicininformationretrievalsystem.Moreover,itdefinesanewpredictioncomputationmethodforthisnewrecommendationmechanism.Experimentalresultsshowthatthealgorithmcanachievebetterpredictionaccuracythantraditionalitem-basedCFalgorithms.Furthermore,thealgorithmcanalleviatethedat
7、asetsparsityproblem.Keywords:E-commerce;recommendationsystem;collaborativefiltering;influenceset摘要:传统的基于用户的协作过滤推荐系统由于使用了基于内存的最近邻查询算法,因此表现出可扩展性差、缺乏稳定性的缺点.针对可扩展性的问题,提出的基于项目的协作过滤算法,仍然不能解决数据稀疏带来的推荐质量下降的问题(稳定性差).从影响集的概念中得到启发,提出一种新的基于项目的协作过滤推荐算法CFBIS(collaborativefilteringbased
8、oninfluencesets),利用当前对象的影响集来提高该资源的评价密度,并为这种新∗SupportedbytheNationalNaturalScienceFoundationofChi
