简介:ThetraditionalGaussianMixtureModel(GMM)forpatternrecognitionisanunsupervisedlearningmethod.Theparametersinthemodelarederivedonlybythetrainingsamplesinoneclasswithouttakingintoaccounttheeffectofsampledistributionsofotherclasses,hence,itsrecognitionaccuracyisnotidealsometimes.ThispaperintroducesanapproachforestimatingtheparametersinGMMinasupervisingway.TheSupervisedLearningGaussianMixtureModel(SLGMM)improvestherecognitionaccuracyoftheGMM.Anexperimentalexamplehasshownitseffectiveness.TheexperimentalresultshaveshownthattherecognitionaccuracyderivedbytheapproachishigherthanthoseobtainedbytheVectorQuantization(VQ)approach,theRadialBasisFunction(RBF)networkmodel,theLearningVectorQuantization(LVQ)approachandtheGMM.Inaddition,thetrainingtimeoftheapproachislessthanthatofMultilayerPerceptrom(MLP).
简介:Recentextensivemeasurementsofreal-lifetrafficdemonstratethattheprobabilitydensityfunctionofthetrafficinnon-Gaussian.Ifatrafficmodeldoesnotcapturethischaracteristics,anyanalyticalorsimulationresultswillnotbeaccurate.Inthiswork,westudytheimpactofnon-Gaussiantrafficonnetworkperformance,andpresentanapproachthatcanaccuratelymodelthemarginaldistributionofreal-lifetraffic.Boththelong-andshort-rangeautocorrelationsarealsoaccounted.Weshowthattheremovalofnon-Gaussiancomponentsoftheprocessdoesnotchangeitscorrelationstructure,andwevalidateourpromisingprocedurebysimulations.
简介:这份报纸为非线性的量的一个班涉及一个过滤问题有多信道的nondemolition大小的随机的系统。系统观察动力学被bosonic的量维纳过程驾驶的随机的微分方程处于真空状态回答的MarkovianHudson-Parthasarathy量管理。作为系统变量的函数,联合操作符的Hamiltonian和系统地被假定在一张Weyl量子化表格被表示。用Wigner-Moyal阶段空间框架,我们为在大小上调节的系统的以后的伪特征功能(QCF)获得一个随机的integro微分的方程。这个方程是与大小联系的革新过程驾驶的Belavkin-Kushner-Stratonovich随机的主人方程的一个空间Fourier领域代表。我们在联合的线性系统地的情况中讨论以后的QCF动力学的一张特定的表格并且构画出以后的量状态的Gaussian近似。