《张子柯-SocialTaggingNetworksStructureDynamics&Applications.ppt》由会员分享,可在线阅读,更多相关《张子柯-SocialTaggingNetworksStructureDynamics&Applications.ppt(18页珍藏版)》请在得力文库 - 分享文档赚钱的网站上搜索。
1、2023/1/21Dep.phys.,Univ.Fribourg1Social Tagging Networks:Structure,Dynamics&ApplicationsCollaborators:Chuang LIU(刘闯),Yi-Cheng ZHANG(张翼成)Tao ZHOU(周涛)Zi-Ke ZHANG(张子柯)Department of Physics,University of Fribourg,SwitzerlandDep.phys.,Univ.Fribourg22023/1/21OutlinenStructure and Dynamics of Social Taggin
2、g NetworksWhat is SNT?Hypergraph StruturesDynamics and emergent propertiesn Applications in Personalized Recommendation(PR)Why Recommendation?How Tags benefit PR?n Conclusions&DiscussionDep.phys.,Univ.Fribourg32023/1/21 What is social tagging networks?Dep.phys.,Univ.Fribourg42023/1/21 What is social
3、 tagging networks?Dep.phys.,Univ.Fribourg52023/1/21From Graph to HypergraphBipartite Network EPL,90(2010)48006Hyper Network from wikipedia.orgUnipartite NetworkDep.phys.,Univ.Fribourg62023/1/21Hypergraph structures in STNZhang and Liu,J.Stat.Mech.(2010)P10005Hyperedge:basic unit Hyper NetworkDep.phy
4、s.,Univ.Fribourg72023/1/21Two roles of social tagsZhang and Liu,J.Stat.Mech.(2010)P10005nRolesRole1:an accessorial tool helping users organize resources:Fig.(a)Role2:a bridge that connects users and resources:Fig.(b)Dep.phys.,Univ.Fribourg82023/1/21Dynamics and evolution of social tagging networks(1
5、/3)nAt each time step,a random user can either:nChoose an item(resource),and annotate it with a relevant or random tag with probability p(Role 1)nor choose a tag,and find a relevant or random item with probability 1-p(Role 2)Dep.phys.,Univ.Fribourg92023/1/21Dynamics of social tagging networks(2/3)nH
6、yperDegree Distribution:nClustering Coefficient:Dep.phys.,Univ.Fribourg102023/1/21Dynamics of social tagging networks(3/3)nAverage Distance:Dep.phys.,Univ.Fribourg112023/1/21nHow to be personalized?Social influence Content-based recommendationNetwork-based recommendation Zhang et al.Physica A 389(20
7、10)179Applications in Personalized Recommendation(PR)nWhy Recommendation?nInformation overload!Dep.phys.,Univ.Fribourg122023/1/21Method I&II:Tripartite Hybrid(Role 1)Zhang et al.Physica A 389(2010)179nItem-user:Method I,PRE 76(2007)046115 nItem-tag:Method IInLinear Hybrid:Dep.phys.,Univ.Fribourg1320
8、23/1/21Method III:Tag-driven(Role 2)Zhang et al.Accepted by EPLnMethod III:Dep.phys.,Univ.Fribourg142023/1/21Algorithm Performance(1/3)Dep.phys.,Univ.Fribourg152023/1/21Algorithm Performance(2/3)Dep.phys.,Univ.Fribourg162023/1/21Algorithm Performance(3/3)Dep.phys.,Univ.Fribourg172023/1/21Conclusions
9、 and DiscussionnConclusionsStructure and DynamicsThe roles of social tagsApplication in Personalized recommendationn DiscussionRecommendation with full hypergraph structureMulti-scale recommendations(semantic-based)Recommendation with community structures2023/1/21Dep.phys.,Univ.Fribourg18Thank You! Zi-Ke ZHANG