trusting spam reporters a reporter-based reputation system for email filtering

《trusting spam reporters a reporter-based reputation system for email filtering》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、TrustingSpamReporters:AReporter-basedReputationSystemforEmailFilteringELENAZHELEVAUniversityofMaryland,CollegeParkALEKSANDERKOLCZMicrosoftLiveLabsLISEGETOORUniversityofMaryland,CollegeParkSpamisagrowingproblem;itinterfereswithvalidemailandburdensbothemailus

2、ersandserviceproviders.Inthiswork,weproposeareactivespam-filteringsystembasedonreporterreputationforuseinconjunctionwithexistingspam-filteringtechniques.Thesystemhasatrust-maintenancecomponentforusers,basedontheirspam-reportingbehavior.Thechallengethatweconsi

3、deristhatofmaintainingareliablesystem,notvulnerabletomalicioususers,thatwillprovideearlyspam-campaigndetectiontoreducethecostsincurredbyusersandsystems.Wereportontheutilityofareputationsystemforspamfilteringthatmakesuseofthefeedbackoftrustworthyusers.Weevalu

4、ateourproposedframework,usingactualcomplaintfeedbackfromalargepopulationofusers,andvalidateitsspam-filteringperformanceonacollectionofrealemailtrafficoverseveralweeks.Totestthebroaderimplicationofthesystem,wecreateamodelofthebehaviorofmaliciousreporters,andwes

5、imulatethesystemundervariousassumptionsusingasyntheticdataset.CategoriesandSubjectDescriptors:H.1[InformationSystems]:ModelsandPrinciplesGeneralTerms:AlgorithmsAdditionalKeyWordsandPhrases:spamfiltering;reputationsystems;trust.Author’saddress:E.Zheleva,Compu

6、terScienceDepartment,AVWilliamsBldg.,UniversityofMaryland,CollegePark,MD,20742.cACM,2009.Thisistheauthor’sversionofthework.ItispostedherebypermissionofACMforyourpersonaluse.Notforredistribution.ThedefinitiveversionwillbepublishedinTOIS,VOL27,ISS1,January2009

7、ACMTransactionsonInformationSystems,Vol.V,No.N,20YY,Pages1–37.2·Zheleva,KolczandGetoor1.INTRODUCTIONEmailspamisagrowingproblemfortheInternetcommunity.Spaminterfereswithvalidemail,anditburdensbothemailusersandemailserviceproviders(ESPs).Itisasourceofannoyanc

8、e,itadverselyaffectsproductivity,andittranslatestosignificantmonetarycostsfortheindustry(e.g.,bandwidth,storage,andthecostofsupportingfilteringinfrastructures).Also,forsomecategoriesofspam,suchasphishscam