文档下载 文档下载导航
设为首页 | 加入收藏
搜索 请输入内容:  
 导航当前位置: 文档下载 > 所有分类 > 高等教育 > 哲学 > A personalized recommendation procedure for Internet shopping support
侵权投诉

A personalized recommendation procedure for Internet shopping support

国外计算机科学类论文,推荐系统

J.K.Kimetal./ElectronicCommerceResearchandApplications1 (2002) 301–313303

onlineretailstore.Ofthesemodels,apartofLeeandPodlaseck’smodelwasadoptedforourresearch,becausetheyfocusontheonlineretailerwhichisalsoourconsideration.They rstanalyzedtheshop-pingbehaviorofcustomersaccordingtothefollow-ingfourshoppingsteps:productimpression,click-through,basketplacement,andpurchase.Theythenappliedmicro-conversionrates(e.g.,click-to-buyrate),computedforeachadjacentpairofthesesteps,tomeasuretheeffectivenessofeffortsinmerchan-dising.

2 .2.Associationrulemining

Givenasetoftransactionswhereeachtransactionisasetofitems(itemset),anassociationruleimpliestheformX Y,whereXandYareitemsets;XandYarecalledthebodyandthehead,respective-ly.ThesupportfortheassociationruleX YisthepercentageoftransactionsthatcontainbothitemsetsXandYamongalltransactions.Thecon dencefortheruleX YisthepercentageoftransactionsthatcontainitemsetYamongtransactionsthatcontainitemsetX[1,13].Thesupportrepresentstheuseful-nessofthediscoveredruleandthecon dencerepresentsthecertaintyoftherule.Associationruleminingisthediscoveryofallassociationrulesthatareaboveauser-speci edminimumsupportandminimumcon dence.

Associationrulemininghasbeenwidelyusedfromtraditionalbusinessapplicationssuchascross-marketing,attachedmailing,catalogdesign,loss-leaderanalysis,storelayout,andcustomersegmenta-tion[1,31],toe-businessapplicationssuchastherenewalofWebpages[10]andWebpersonalization[22,24].

Associationruleminingcanbeperformedatmultiplelevelsofhierarchiesovertheitemduetoitsvaluableusageasindicatedbymanyresearchers[11–13,18,31,33].Rulesgeneratedfromassociationruleminingwithhierarchiesarecalledmultiple-levelormulti-levelassociationrules,sincetheyconsidermorethanonelevelofhierarchies.

2 .3.Producttaxonomy

Inmostonlinestores,aproducttaxonomyisavailable.Aproducttaxonomyispracticallyrepre-sentedasatreethatclassi esasetoflow-level

productsintoahigherlevelofmoregeneralprod-ucts.Theleavesofthetreedenotetheproductinstances,SKUs(StockKeepingUnits)inretailjargon,andnon-leafnodesdenoteproductclassesobtainedbycombiningseverallower-levelnodesintooneparentnode.TherootnodelabeledAlldenotesthemostgeneralproductclass.Fig.1showsanexampleproducttaxonomyfora ctionalbeautyInternetshoppingmall,whereConditioner,PomadeandShampooareclassi edintoHaircare,etc.Anumbercalledlevelcanbeassignedtoeachnodeintheproducttaxonomy.Theleveloftherootnodeiszero,andthelevelofanyothernodeisoneplusthelevelofitsparent.Notethatahigher-levelproductclasshasasmallerlevelnumber.TheproducttaxonomyofFig.1has velevels,referredtoaslevel0(forroot),1,2,3,and4.

Producttaxonomiesplayanimportantroleintheknowledgediscoveryprocesssincetheyrepresentonlinestoredependentknowledgeandmayaffecttheresults.Theyareusedforidentifyingsimilarprod-uctsandgroupingthemtogether,byspecifyingthelevelofaggregationprovidedbythemarketersordomainexperts.Forinstance,miningassociationrulesatdifferentlevelsoftheproducttaxonomyarereportedtoresultinthediscoveryofmorespeci candconcreteknowledgeratherthanatasinglelevel[11].Furthermore,itisusefultohaveataxonomyoftheproductsbeingconsideredfordatamininganaly-sisbecausechoosingtherightlevelofproducttaxonomymayimprovetheresultsoftheanalysis[5].

2 .4.Decisiontreeinduction

Predictivedataminingisasearchforverystrongpatternsinalargedatasetthatcangeneralizeaccuratefuturedecisions.Decisiontreeinductiontechniquesarethemostwidelyusedpredictionmethodsandrelativelysmalltreesareeasytounderstand.Decisiontreeinductiontechniquesbuilddecisiontreestolabelorcategorizecasesintoasetofknownclasses.Incontrasttoneuralnetworks,decisiontreeinductiontechniquesarewellsuitedforhigh-dimensionalapplicationsandhavestrongexpla-nationcapabilities[4,17].

Inaninduceddecisiontree,eachnon-leafnodedenotesatestonanattribute,eachbranchcorre-

上一页第3页

你可能喜欢

热门文档

相关文档

站点地图 | 文档上传 | 侵权投诉 | 手机版
本站所有资源均来自互联网,本站只负责收集和整理,均不承担任何法律责任,如有侵权等其它行为请联系我们.
文档下载 Copyright 2013 doc.guandang.net All Rights Reserved.  -  email