简介：

简介：Inordertoimprovetheabilitytolocalizeasourceinanuncertainacousticenvironment,aBayesianapproach,referredtohereasBayesianlocalizationisusedbyincludingtheenvironmentintheparametersearchspace.Geneticalgorithmsareusedfortheparameteroptimization.Thismethodintegratestheaposteriorprobabilitydensity(PPD)overenvironmentalparameterstoobtainasequenceofmarginalprobabilitydistributionsoversourcerangeanddepth,fromwhichthemost-probablesourcelocationandlocalizationuncertaintiescanbeextracted.Consideringthattheseabeddensityandattenuationarelesssensitivetotheobjectivefunctionofmatchedfieldprocessing,weutilizetheempiricalrelationshiptoinvertthoseparametersindirectly.ThebroadbandsignalsrecordedbyaverticallinearrayinaYellowSeaexperimentin2000areprocessedandanalyzed.Itwasfoundthat,theBayesianlocalizationmethodthatincorporatestheenvironmentalvariabilityintotheprocessor,madeitrobusttotheuncertaintyintheoceanenvironment.Inaddition,usingtheempiricalrelationshipcouldenhancethelocalizationaccuracy.

简介：

简介：

简介：

简介：Abundanttestdataarerequiredinassessmentofweaponperformance.Whenweapontestdataareinsufficient,Bayesiananalysesinsmallsamplecircumstanceshouldbeconsideredandthetestdatashouldbeprovidedbysimulations.TheseveralBayesianapproachesarediscussedandsomelimitationsarefounded.AnimprovementisputforwardafterlimitationsofBayesianapproachesavailableareanalyzedandtheimprovedapproachisappliedtoassessmentofsomenewweaponperformance.

简介：

简介：ThelearningBayesiannetwork(BN)structurefromdataisanNP-hardproblemandstilloneofthemostexcitingchallengesinthemachinelearning.Inthiswork,anovelalgorithmispresentedwhichcombinesideasfromlocallearning,constraintbased,andsearch-and-scoretechniquesinaprincipledandeffectiveway.ItfirstreconstructsthejunctiontreeofaBNandthenperformsaK2-scoringgreedysearchtoorientatethelocaledgesinthecliquesofjunctiontree.Theoreticalandexperimentalresultsshowtheproposedalgorithmiscapableofhandlingnetworkswithalargenumberofvariables.Itscomparisonwiththewell-knownK2algorithmisalsopresented.

简介：WordSenseDisambiguation(WSD)istodecidethesenseofanambiguouswordonparticularcontext.MostofcurrentstudiesonWSDonlyuseseveralambiguouswordsastestsamples,thusleadstosomelimitationinpracticalapplication.Inthispaper,weperformWSDstudybasedonlargescalereal-worldcorpususingtwounsupervisedlearningalgorithmsbasedon±n-improvedBayesianmodelandDependencyGrammar(DG)-improvedBayesianmodel.±n-improvedclassifiersreducethewindowsizeofcontextofambiguouswordswithclose-distancefeatureextractionmethod,anddecreasethejammingofuselessfeatures,thusobviouslyimprovetheaccuracy,reaching83.18%(inopentest).DG-improvedclassifiercanmoreeffectivelyconquerthenoiseeffectexistinginNaive-Bayesianclassifier.ExperimentalresultsshowthatthisapproachdoesbetteronChineseWSD,andtheopentestachievedanaccuracyof86.27%.

简介：Recently,variableselectionbasedonpenalizedregressionmethodshasreceivedagreatdealofattention,mostlythroughfrequentist'smodels.ThispaperinvestigatesregularizationregressionfromBayesianperspective.OurnewmethodextendstheBayesianLassoregression(ParkandCasella,2008)throughreplacingtheleastsquarelossandLassopenaltybycompositequantilelossfunctionandadaptiveLassopenalty,whichallowsdifferentpenalizationparametersfordifferentregressioncoefficients.BasedontheBayesianhierarchicalmodelframework,anefficientGibbssamplerisderivedtosimulatetheparametersfromposteriordistributions.Furthermore,westudytheBayesiancompositequantileregressionwithadaptivegroupLassopenalty.Thedistinguishingcharacteristicofthenewlyproposedmethodiscompletelydataadaptivewithoutrequiringpriorknowledgeoftheerrordistribution.Extensivesimulationsandtworealdataexamplesareusedtoexaminethegoodperformanceoftheproposedmethod.Allresultsconfirmthatournovelmethodhasbothrobustnessandhighefficiencyandoftenoutperformsotherapproaches.

简介：调查早lycopsids，用吝啬的cladistic分析和贝叶斯的途径的发展史被介绍，与越过25taxa的33个词法人物的一个数据矩阵。产生吝啬和贝叶斯的树在树拓扑学显示出全面类似。在贝叶斯的树上，Protolepidodendrales和heterosporouslycopsids作为二个monophyletic组被认出，并且在后者以内组织Isoetalessensulato形式subclade。这拓扑学暗示一个舌状的人物的获得在lycopsids演变两次在protolepidodendralean雷克勒基亚·班克斯等，并且一次在heterosporouslycopsidclade。以前从华南的中间上面的泥盆纪报导的几lycopsid植物的种系发生的位置，有不明确的顺序的亲密关系，被cladistic分析估计；cf。薛伦戈斯塔奇斯etHao，朱伦戈斯塔奇斯等，王莫尼利斯特罗布斯et果浆，和YuguangiaHao等。在Isoetalesclade以内掉落很好，作为这份订单的最早的成员，当时王米诺斯特罗布斯和Wuxia果浆等。在heterosporouslycopsids的基础部分以内被嵌套。贝叶斯的分析是在系统的研究的一条很有用的途径并且能在分析数据也设置的paleobotanical被使用。

简介：ABayesianestimationmethodtoseparatemulticomponentsignalswithsinglechannelobservationispresentedinthispaper.Byusingthebasisfunctionprojection,thecomponentseparationbecomesaproblemoflimitedparameterestimation.Then,aBayesianmodelforestimatingparametersissetup.ThereversiblejumpMCMC(MonteCarloMarkovChain)algorithmisadoptedtoperformtheBayesiancomputation.Themethodcanjointlyestimatetheparametersofeachcomponentandthecomponentnumber.SimulationresultsdemonstratethatthemethodhaslowSNRthresholdandbetterperformance.

简介：

简介：1.IntroductionAnimportsnttopicinQualityColltrolisthestudyofacceptancesamplingplan.Foravariablesamplingplan,thequalityofanitemismeasuredbyarandomvariable.Todesignavariablesamplingplanistodeterminethesamplesizeandthespecilicstionlicit(s).Manyschemeshav...

简介：贝叶斯的匹配的地处理的一条途径(MFP)在不明确的海洋environment.In被讨论这条途径，无常知识被建模，数组收到的空间、时间的数据是充分used.Therefore，为MFP的机制是found.which很好联合不明确的领域processing.By的基于模型、数据驱动的方法理论推导，模拟分析和在海的试验性的数组数据的确认，我们发现那(1)贝叶斯的匹配的领域pr的基本部件

简介：AnewmethodalongwithBayesianapproachforestimatingtheparameterinthedistributionfunctionF(x;θ)byusinggroupeddataisdevelopedinthispaper.ThesupportofF(x;θ)isdividedintodisjointedintervalsas-∞=x0

简介：公共天气服务是向向用户提供概率的天气预报的trending，代替传统的确定的预报。概率的预报技术不断地正在被改进优化可得到的预报信息。预报(BPF)的贝叶斯的处理器，为概率的预报的一个新统计方法，能根据在那个预报系统产生的观察和预报之间的历史的统计关系把一张确定的预报转变成一张概率的预报。这种技术在确定说明一个确定的预报系统的典型预报性能预报无常。meta-Gaussian可能性的模型对有单调可能性的比率的许多随机的依赖结构合适。收养这种可能性的模特儿的meta-GaussianBPF能因此越过许多地被使用，包括气象学和水文学。有二个连续随机的变量和正常线性的BPF的Bayes定理简短被介绍。为用一个单个预言者的连续predictand的meta-GaussianBPF然后被介绍并且讨论。meta-GaussianBPF的表演在一个初步的实验被测试。在在长沙和武汉车站的0000UTC的每日的表面温度的控制预报被用作确定的预报数据。这些控制预报从整体预言被拿，一96-h铅时间由中国气象学的管理的国家气象学的中心产生了，中等范围的天气的欧洲中心预报，并且US公民为在2008年1月期间的环境预言集中。实验的结果证明meta-GaussianBPF能从三整体预言中的任何一个把表面温度的一张确定的控制预报转变成表面温度的一张有用概率的预报。这些概率的预报确定控制预报的无常；因此，概率的预报的表演基于内在的确定的控制预报的来源不同。

简介：

简介：TheBayesianapproachisconsideredasthemostgeneralformulationofthestateestimationfordynamicsystems.However,mostoftheexistingBayesianestimatorsofstochastichybridsystemsonlyfocusontheMarkovjumpsystem,fewliteratureisrelatedtotheestimationproblemofnonlinearstochastichybridsystemswithstatedependenttransitions.Accordingtothisproblem,anewmethodologywhichrelaxesquitearestrictiveassumptionthatthemodetransitionprocessmustsatisfyMarkovpropertiesisproposed.Inthismethod,ageneralapproachispresentedtomodelthestatedependenttransitions,thestateandoutputspacesarediscretedintocellspacewhichhandlesthenonlinearitiesandcomputationallyintensiveproblemoffline.ThenmaximumaposteriorestimationisobtainedbyusingtheBayesiantheory.Theefficacyoftheestimatorisillustratedbyasimulatedexample.

简介：OrderingbasedsearchmethodshaveadvantagesovergraphbasedsearchmethodsforstructurelearningofBayesiannetworksintermsontheefficiency.Withtheaimoffurtherincreasingtheaccuracyoforderingbasedsearchmethods,wefirstproposetoincreasethesearchspace,whichcanfacilitateescapingfromthelocaloptima.Wepresentoursearchoperatorswithmajorizations,whichareeasytoimplement.Experimentsshowthattheproposedalgorithmcanobtainsignificantlymoreaccurateresults.Withregardtotheproblemofthedecreaseonefficiencyduetotheincreaseofthesearchspace,wethenproposetoaddpathpriorsasconstraintsintotheswapprocess.Weanalyzethecoefficientwhichmayinfluencetheperformanceoftheproposedalgorithm,theexperimentsshowthattheconstraintscanenhancetheefficiencygreatly,whilehaslittleeffectontheaccuracy.Thefinalexperimentsshowthat,comparedtoothercompetitivemethods,theproposedalgorithmcanfindbettersolutionswhileholdinghighefficiencyatthesametimeonbothsyntheticandrealdatasets.