rlm a general model for trust representation and aggregation

rlm a general model for trust representation and aggregation

ID:7291855

大小:3.06 MB

页数:14页

时间:2018-02-10

rlm a general model for trust representation and aggregation_第1页
rlm a general model for trust representation and aggregation_第2页
rlm a general model for trust representation and aggregation_第3页
rlm a general model for trust representation and aggregation_第4页
rlm a general model for trust representation and aggregation_第5页
资源描述:

《rlm a general model for trust representation and aggregation》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、IEEETRANSACTIONSONSERVICESCOMPUTING,VOL.X,NO.X,XXX20101RLM:AGeneralModelforTrustRepresentationandAggregationXiaofengWang,Member,IEEE,LingLiu,SeniorMember,IEEEandJinshuSu,Member,IEEEAbstract—Reputation-basedtrustsystemsprovideimportantcapabilityinopenandservice-oriented

2、computingenvironments.Mostexistingtrustmodelsfailtoassessthevarianceofareputationprediction.Moreover,thesummationmethod,widelyusedforreputationfeedbackaggregation,isvulnerabletomaliciousfeedbacks.Thispaperpresentsageneraltrustmodel,calledRLM,foramorecomprehensiveandrob

3、ustreputationevaluation.Concretely,wedefineacomprehensivereputationevaluationmethodbasedontwoattributes:reputationvalueandreputationpredictionvariance.Thereputationpredicationvarianceservesasaqualitymeasureofthereputationvaluecomputedbasedonaggregationoffeedbacks.Forfee

4、dbackaggregation,weproposethenovelKalmanaggregationmethod,whichcaninherentlysupportrobusttrustevaluation.Todefendagainstmaliciousandcoordinatedfeedbacks,wedesigntheExpectationMaximizationalgorithmtoautonomouslymitigatetheinfluenceofamaliciousfeedback,andfurtherapplytheh

5、ypothesistestmethodtoresistmaliciousfeedbacksprecisely.Throughtheoreticalanalysis,wedemonstratetherobustnessoftheRLMdesignagainstadulatinganddefamingattacks,twopopulartypesoffeedbackattacks.OurexperimentsshowthattheRLMmodelcaneffectivelycapturethereputation’sevolutiona

6、ndoutperformthepopularsummationbasedtrustmodelsintermsofbothaccuracyandattackresilience.Concretely,undertheattackofcollusivemaliciousfeedbacks,RLMoffershigherrobustnessforthereputationpredictionandalowerfalsepositiverateforthemaliciousfeedbackdetection.IndexTerms—trust

7、model,accuracyassessment,maliciousfeedback,robustness.F1INTRODUCTIONrectlybetweenprovidersandtheevaluator(personalexperience)andtherecommendationsmadebyotherTherapidgrowthofInternetandubiquitousconnectiv-consumers(feedback)[1].Fromthepersonalexperi-ityhasspurredthedeve

8、lopmentofvariouscollaborativeence’sperspective,mostexistingworkusedthesimplecomputingsystemssuchasservice-orientedcom

当前文档最多预览五页,下载文档查看全文

此文档下载收益归作者所有

当前文档最多预览五页,下载文档查看全文
温馨提示:
1. 部分包含数学公式或PPT动画的文件,查看预览时可能会显示错乱或异常,文件下载后无此问题,请放心下载。
2. 本文档由用户上传,版权归属用户,天天文库负责整理代发布。如果您对本文档版权有争议请及时联系客服。
3. 下载前请仔细阅读文档内容,确认文档内容符合您的需求后进行下载,若出现内容与标题不符可向本站投诉处理。
4. 下载文档时可能由于网络波动等原因无法下载或下载错误,付费完成后未能成功下载的用户请联系客服处理。