简介:Becauseofitswideapplicationinanonymousauthenticationandattribute-basedmessaging,theattribute-basedsignatureschemehasattractedthepublicattentionsinceitwasproposedin2008.However,mostoftheexistingattribute-basedsignatureschemesarenolongersecureinquantumera.Fortunately,lattice-basedcryptographyoffersthehopeofwithstandingquantumcomputers.Andlatticeshaselevatedittothestatusofapromisingpotentialalternativetocryptographybasedondiscretelogandfactoring,owingtoimplementationsimplicity,provablesecurityreductionsandquantum-immune.Inthispaper,thefirstlatticeattribute-basedsignatureschemeinrandomoraclemodelisproposed,whichisprovedexistentialunforgeabilityandperfectprivacy.Comparedwiththecurrentattribute-basedsignatureschemes,ournewattribute-basedsignatureschemecanresistquantumattacksandhasmuchshorterpublic-keysizeandsignaturesize.Furthermore,thisschemeisextendedintoanattribute-basedsignatureschemeonnumbertheoryresearchunit(NTRU)lattice,whichisalsosecureeveninquantumeraandhasmuchhigherefficiencythantheformer.
简介:在许多机器学习应用程序,数据不是免费的,并且为每个数据项目有测试费用。为节俭的原因,最小化测试的某存在工作尝试花费了并且同时,保存一个给定的决定系统的一个特别性质。在这份报纸,我们指出一个人能负担得起的测试费用在一些应用是有限的。因此,一个人不得不牺牲各自的性质在一项预算下面保留测试费用。形式化这期,我们定义测试费用限制属性减小问题,在优化目的是最小化有条件的信息熵的地方。这个问题是test-cost-sensitive属性减小问题和0-1背囊问题的必要归纳,因此它是更挑战性的。我们基于信息获得和测试费用建议一个启发式的算法处理新问题。算法在四UCI上被测试(加利福尼亚大学-Irvine)有各种各样的测试的数据集花费了背景。试验性的结果显示唯一的userspecified参数的适当设置。
简介:Anapproachtosupportinguserinterfacesusinganattributegrammarcombinedwithaneventmodelisdescribed.Themainemphasisishowtorepresentamulti-threaddialoguemodelindirectmanipulationuserinterfaces.Itisshownthatcontrolsequencewithindialogues,communicationwithotherdialogues,andsomecomputationsforapplicationscanbespecifiedwithasyntacticandsemanticnotation.Theattributegrammarspecificationcanbeimplementedusinganattributegrammarintepreterembeddedintherun-timestructuresupportingcommunication,synchronization,anddialogueexecuton.
简介:特征选择(FS)目的将当在代表原来的特征保留合适高的精确性时,从一个问题领域决定一个最小的特征(属性)子集。当有许多success.RST的如此的一个工具启用数据相关性和独自用数据在数据集包含的属性的数字的减小的发现,不平的集合理论(RST)被使用了,不要求另外的信息。这份报纸在基于RST的途径后面描述基本想法,考察在这些想法上造的相关FS方法,并且分析象Quickreduct算法那样的更经常使用的基于RST的传统的FS算法,熵基于编校算法,和相对编校算法。一些在存在算法和我们的建议改进算法罐头的缺点克服这些缺点,这被发现。试验性的分析被执行了以便完成建议算法的效率。
简介:Inthispaperanewsignaturescheme,calledPolicy-EndorsingAttribute-BasedSignature,isdevelopedtocorrespondwiththeexistingCiphertext-PolicyAttribute-BasedEncryption.Thissignatureprovidesapolicy-and-endorsementmechanism.Inthismechanismasingleuser,whoseattributessatisfythepredicate,endorsesthemessage.Thissignatureallowsthesignertoannouncehisendorsementusinganaccesspolicywithouthavingtorevealtheidentityofthesigner.Thesecurityofthissignature,selflessanonymityandexistentialunforgeability,isbasedontheStrongDiffie-HellmanassumptionandtheDecisionLinearassumptioninbilinearmapgroups.
简介:Toeliminatecomputationalproblemsinvolvedinevaluatingmulti-attributebidswithdifferentmeasures,thisarticlefirstnormalizeseachindividualcomponentofabid,andthenmakesuseoftheweightedproductmethodtopresentanewscoringfunctionthatconvertseachbidintoascore.Twokindsofmulti-attributeauctionmodelsareintroducedintermsofscoringrulesandbiddingobjectivefunctions.Equilibriumbiddingstrategies,procurer'srevenuecomparisonsandoptimalauctiondesignarecharacterizedinthesetwomodels.Finally,thisarticlediscussessomeimprovementofrobustnessofourmodels,withrespecttotheassumptions.
简介:Onthebasisoftheprinciplesofsimplerandomsampling,thestatisticalmodelofrateofdisfigurement(RD)isputforwardanddescribedindetail.AccordingtothedefinitionofsimplerandomsamplingfortheattributedatainGIS,themeanandvarianceoftheRDarededucedasthecharacteristicvalueofthestatisticalmodelinordertoexplainthefeasibilityoftheaccuracymeasurementoftheattributedatainGISbyusingtheRD.Moreover,onthebasisofthemeanandvarianceoftheRD,thequalityassessmentmethodforattributedataofvectormapsduringthedatacollectingisdiscussed.TheRDspreadgraphisalsodrawntoseewhetherthequalityoftheattributedataisundercontrol.TheRDmodelcansyntheticallyjudgethequalityofattributedata,whichisdifferentfromothermeasurementcoefficientsthatonlydiscussaccuracyofclassification.
简介:到摘录和快车,在信息系统,discernibility矩阵和它的扩展隐藏的知识在许多真实生活应用程序成功地被介绍给并且适用。作为一条代表性的途径,二进制discernibility矩阵有许多有趣的优异性质并且很快被开发了发现直觉、容易理解知识。目前,然而,二进制discernibility矩阵主要在完全的信息系统被采用。怎么由在不完全的信息系统使用二进制discernibility矩阵完成属性减小是一个挑战性的话题。概括二进制discernibility矩阵的一种形式进一步为处理不完全的信息系统的很多个代表性的扩大不平的集合模型被开发。一些有用性质和标准为判定属性核心和属性亲戚减小被介绍。此后,一个新算法被提出它支持基于概括二进制discernibility矩阵归因核心和属性亲戚减小。这个算法对一致信息系统而且不一致的信息系统合适不仅。建议方法的可行性被工作例子和试验性的分析表明。
简介:Particleswarmoptimization(PSO)isanewheuristicalgorithmwhichhasbeenappliedtomanyoptimizationproblemssuccessfully.Attributereductionisakeystudyingpointoftheroughsettheory,andithasbeenproventhatcomputingminimalreductionofdecisiontablesisanon-derterministicpolynomial(NP)-hardproblem.AnewcooperativeextendedattributereductionalgorithmnamedCo-PSARbasedonimprovedPSOisproposed,inwhichthecooperativeevolutionarystrategywithsuitablefitnessfunctionsisinvolvedtolearnagoodhypothesisforacceleratingtheoptimizationofsearchingminimalattributereduction.ExperimentsonBenchmarkfunctionsandUniversityofCalifornia,Irvine(UCI)datasets,comparedwithotheralgorithms,verifythesuperiorityoftheCo-PSARalgorithmintermsoftheconvergencespeed,efficiencyandaccuracyfortheattributereduction.
简介:Bycombiningtheadvantagesoftheadditiveweightedmean(AWM)operatorandtheorderedweightedaveraging(OWA)operator,thispaperfirstpresentsahybridoperatorforaggregatingdatainformation,andthenproposesahybridaggregation(HA)operator-basedmethodformultipleattributedecisionmaking(MADM)problems.ThetheoreticalanalysesandthenumericalresultsshowthattheHAoperatorgeneralizesboththeAWMandOWAoperators,andreflectstheimportanceofboththegivenargumentandtheorderedpositionoftheargument.Thus,theHAoperatorcanreflectbetterrealsituationsinpracticalapplications.Finally,anillustrativeexampleisgiven.