简介:Thispaperpresentsanoveltimedelayestimation(TDE)methodusingtheconceptofentropy.Therelativedelayisestimatedbyminimizingtheestimatedjointentropyofmultiplesensoroutputsignals.Whenestimatingtheentropy,theinformationaboutthepriordistributionofthesourcesignalisnotrequired.Instead,theParzenwindowestimatorisemployedtoestimatethedensityfunctionofthesourcesignalfrommultiplesensoroutputsignals.Meanwhile,basedontheParzenwindowestimator,theRenyi’squadraticentropy(RQE)isincorporatedtoeffectivelyandefficientlyestimatethehigh-dimensionaljointentropyofthemultichanneloutputs.Furthermore,amodifiedformofthejointentropyforembeddinginformationaboutreverberation(multipathreflections)forspeechsignalsisintroducedtoenhancetheestimator’srobustnessagainstreverberation.
简介:地震小浪评价是地震数据处理和解释的重要部分,谁的正确直接与deconvolution和倒置的结果有关。小浪评价基于高顺序的系列是一个重要新方法。然而,高顺序的系列经常有阶段包装纸问题,它导致小浪阶段光谱偏差并且从而影响混合阶段小浪评价。解决这个问题,我们建议一个新阶段光谱方法基于在bispectral领域的保角的印射。方法避免由缩小Fourier阶段光谱的范围消除在原来的阶段包影响的bispectral阶段包问题的阶段光谱评价。方法基于保角的印射组成最少平方的小浪阶段光谱评价它与最少平方的小浪振幅光谱评价被用于混合阶段小浪评价。理论模型和实际地震数据验证这个方法的有效性。我们也扩大到trispectral小浪阶段光谱评价的在bispectral小浪阶段光谱评价的保角的印射的想法。
简介:Inthispaper,anewtimingoffsetestimationmethodforOrthogonalFrequencyDivisionMultiplexing(OFDM)systemsarepresented.Uponthereceiptofjustonetrainingsequence,thispresenceofasignalandthestartoftheframecanbedetected.TheperformancesofthetimingestimatorinadditivewhiteGaussiannoiseandthemobilemultipathchannelareevaluatedbysimulation.TheresultsdeliveredbythisalgorithmarecomparedwiththosedeliveredbySchmidlandCox.
简介:TheMTD(mixturetransitiondistribution)modelbasedonWeibulldistribution(WMTDmodel)isproposedinthispaper,whichisaimedatitsparameterestimation.AnEMalgorithmforestimationisgivenandshowntoworkwellbysomesimulations.Andbootstrapmethodisusedtoobtainconfidenceregionsfortheparameters.Finally,theresultsofarealexample-predictingstockprices-showthattheWMTDmodelproposedisabletocapturethefeaturesofthedatafromthick-taileddistributionbetterthanGMTD(mixturetransitiondistribution)model.
简介:Handlingindustrialpowdersleadstodustemissions.Theseemissionstendtogeneratehumandiseasesorcauseotherenvironmentaleffects.Amultitudeofapparatushasbeendevelopedtoestimatethisdustinessbehaviourofpowders.Twoofthesewell-knownmethodsarepresentedandcomparedwitheachother.Athirdrecentlydevelopedmethodisalsointroduced.
简介:Bargraphsareconvenientforshowingcomparisonsamongitems.Thebarsmaybeeitherhorizontalorvertical,andtheyareusedtoshowtheamountsofdifferentitems.Figure1isanexampleofatypicalbargraph.ThegraphshowstheaveragesalespriceofexistinghomesintheNortheastUnitedStatesforthreeyears——1970,1980,and1997.
简介:Inthispaper,admissibilityandinadmissibilityresultsofsimultaneousestimationoflossesofmultivariatenormalmeanestimatorsaregiven.Admissibleestimatorsforlossesofusualestimatorsofmultivariategammaparameters,multivariatebinomialparametersandPoissonparametersareobtainedrespectively.
简介:Researchonjumpregressionfunc(?)onshasnotbeenadequateyetAccordingtotheinformationaboutthenumberofjumps,theirpositionsandjumpmagnitudes,jumpregressionfunctionscanbeclassifiedintoeighttypes.Thispaperdealsespeciallywiththesecondjumpregressionfunction.Firstofall,aconceptoftrimmedsplineestimateisproposedandwithitanL~2-consistentestimateofthesmoothingpartofthejumpregressionfunctionisobtained.ThisalongwiththeL~2-consistentestimateofjumpmagnitudeconstitutesanestimateofthesecondjumpregressionfunction.Thispaperdiscussesalsothecasethatthejumppositionshavesomeindeterminacy.Anewcriterionissuggestedanditsuniquesolutionderived.Intheend,afewnumericalresultsaregiven.
简介:Basedonuniformcirculararrayanditselementoutputsignaldelaydelayanovelmethodforestimationofspatialfrequency,azimuthandelevationispresented.Withoutanyspectralpeaksearchandparameterspairing,theresultingmethodis(?)resolutionandsmallvariance,eveninshortdatalength.Computersimulation(?)effectivenessofthismethod.
简介:Inthisnoteweshowthatforagivencontrollablepair(A,B)andanyλ>0,againmatrixKcanbechosensothatthetransitionmatrixe(A+BK)tofthesystemx=(A+BK)xdecaysattheexponentialratee-λtandtheovershootofthetransitionmatrixcanbeboundedbyMλLforsomeconstantsMandLthatareindependentofλ.Asaconsequence,foranyh>0,againmatrixKcanbechosensothatthemagnitudeofthetransitionmatrixe(A+BK)tcanbereducedby-1/2-(orbyanygivenportion)over[0,h].Aninterestingapplicafonoftheresultisinthestabilizationofswitchedlinearsystemswithanygivenswitchingrate(see[1]).
简介:Inapplications,thetraditionalestimationproceduregenerallybeginswithmodelselection.Onceaspecificmodelisselected,subsequentestimationisconductedundertheselectedmodelwithoutconsiderationoftheuncertaintyfromtheselectionprocess.Thisoftenleadstotheunderreportingofvariabilityandtoooptimisticconfidencesets.Modelaveragingestimationisanalternativetothisprocedure,whichincorporatesmodeluncertaintyintotheestimationprocess.Inrecentyears,therehasbeenarisinginterestinmodelaveragingfromthefrequentistperspective,andsomeimportantprogresseshavebeenmade.Inthispaper,thetheoryandmethodsonfrequentistmodelaveragingestimationaresurveyed.Somefutureresearchtopicsarealsodiscussed.
简介:AnewDirectionOfArrival(DOA)estimationalgorithmforwidebandsourcesbasedonUniformCircularArray(UCA)ispresentedviaanalyzingwidebandperformanceofthegeneralESPRIT.ThealgorithmeffectivelyimprovesthewidebandperformanceofESPRITbasedontheinterpolationprincipiumandUCA-ESPRIT.Thesimulatedresultsbycomputerdemonstrateitsefficiency.
简介:Inordertogiveacompleteandaccuratedescriptionaboutthesensitivityofefficientportfoliostochangesinassets'expectedreturns,variancesandcovariances,thejointeffectofestimationerrorsinmeans,variancesandcovariancesontheefficientportfolio'sweightsisinvestigatedinthispaper.Itisprovedthattheefficientportfolio'scompositionisaLipschitzcontinuous,differentiablemappingoftheseparametersundersuitableconditions.Thechangerateoftheefficientportfolio'sweightswithrespecttovariationsaboutriskreturnestimationsisderivedbyestimatingtheLipschitzconstant.Ourgeneralquantitativeresultsshowthattheefficientportfolio''sweightsarenormallynotsosensitivetoestimationerrorsaboutmeansandvariances.Moreover,wepointoutthoseextremecaseswhichmightcausestabilityproblemsandhowtoavoidtheminpractice.Preliminarynumericalresultsarealsoprovidedasanillustrationtoourtheoreticalresults.