資源描述:
《an analysis of the exponential decay principle in probabilistic》由會員上傳分享,免費在線閱讀,更多相關(guān)內(nèi)容在工程資料-天天文庫。
1、TheoreticalComputerScience410(2009)40674084ContentslistsavailableatScienceDirectTheoreticalComputerSciencejournalhomepage:www.elsevier.com/locate/tcsAnanalysisoftheexponentialdecayprincipleinprobabilistictrustmodelsEhabElSalamounya,KarlTikj?bKrukowb,VladimiroSass
2、onea,aECS,UniversityofSouthampton,UKbTrifork,Aarhus,DenmarkarticleinfoabstractResearchinmodelsforexperience-basedtrustmanagementhaseitherignoredtheproblemThispaperisdedicatedtoProfessorMogensofmodellingandreasoningaboutdynamicallychangingprincipalbehaviour,orprov
3、idedNielsenontheoccasionofhis60thadhocsolutionstoit.Probabilitytheoryprovidesafoundationforaddressingthisandbirthday,whoalongtheyearssharedwithmanyotherissuesinarigorousandmathematicallysoundmanner.UsingHiddenMarkovusresearchandmuchmore.ToyourModelstorepresentprin
4、cipalbehaviours,wefocusoncomputationaltrustframeworkssharpnessofmindandkindnessofspirit.basedonthe`beta'probabilitydistributionandtheprincipleofexponentialdecay,andKeywords:deriveapreciseanalyticalformulafortheestimationerrortheyinduce.ThisallowsTrustpotentialadop
5、tersofbeta-basedcomputationaltrustframeworksandalgorithmstobetterComputationaltrustunderstandtheimplicationsoftheirchoice.Probabilistictrustmodels'2009ElsevierB.V.Allrightsreserved.HiddenMarkovModelsDecayprincipleBetadistribution1.IntroductionThispaperconcernsexpe
6、rience-basedtrustmanagementsystems.Theterm`trustmanagement'isusuallyassociatedwiththetraditionalcredential-basedtrustmanagementsystemsinwhichtrustisestablishedprimarilyasafunctionofavailablecredentials(seee.g.[2]).Inexperience-basedsystems,trustinaprincipalisrepre
7、sentedasafunctionofinformationaboutpastprincipalbehaviour.Thisencompassesalsoreputationinformation,i.e.,informationaboutprincipalbehaviourobtainednotbydirectobservationbutfromothersources(e.g.,ratingsmadebyotherprincipals).Thereareseveralapproachestoexperience-bas
8、edtrustmanagement;thispaperisconcernedwiththeprobabilisticapproach,whichcanbroadlybecharacterisedasaimingtobuildprobabilisticmodelsuponwhichtobasepredic