Table Of ContentDigitalSignalProcessingwithKernelMethods
Digital Signal Processing with Kernel Methods
JoséLuisRojo-Álvarez
DepartmentofSignalTheoryandCommunications
UniversityReyJuanCarlos
Fuenlabrada(Madrid)
and
CenterforComputationalSimulation
UniversidadPolitécnicadeMadrid,Spain
ManelMartínez-Ramón
DepartmentofElectricalandComputerEngineering
TheUniversityofNewMexico
Albuquerque,NewMexico
USA
JordiMuñoz-Marí
DepartmentofElectronicsEngineering
UniversitatdeValència
Paterna(València),Spain
GustauCamps-Valls
DepartmentofElectronicsEngineering
UniversitatdeValència
Paterna(València),Spain
Thiseditionfirstpublished
©JohnWiley&SonsLtd
Allrightsreserved.Nopartofthispublicationmaybereproduced,storedinaretrievalsystem,or
transmitted,inanyformorbyanymeans,electronic,mechanical,photocopying,recordingorotherwise,
exceptaspermittedbylaw.Adviceonhowtoobtainpermissiontoreusematerialfromthistitleisavailable
athttp://www.wiley.com/go/permissions.
TherightofJoséLuisRojo-Álvarez,ManelMartínez-Ramón,JordiMuñoz-Marí,GustauCamps-Vallstobe
identifiedastheauthorsoftheeditorialmaterialinthisworkhasbeenassertedinaccordancewithlaw.
RegisteredOffices
JohnWiley&Sons,Inc.,RiverStreet,Hoboken,NJ,USA
JohnWiley&SonsLtd,TheAtrium,SouthernGate,Chichester,WestSussex,POSQ,UK
EditorialOffice
TheAtrium,SouthernGate,Chichester,WestSussex,POSQ,UK
Fordetailsofourglobaleditorialoffices,customerservices,andmoreinformationaboutWileyproducts
visitusatwww.wiley.com.
Wileyalsopublishesitsbooksinavarietyofelectronicformatsandbyprint-on-demand.Somecontentthat
appearsinstandardprintversionsofthisbookmaynotbeavailableinotherformats.
LimitofLiability/DisclaimerofWarranty
MATLABⓇandSimulinkisatrademarkofTheMathWorks,Inc.andisusedwithpermission.The
MathWorksdoesnotwarranttheaccuracyofthetextorexercisesinthisbook.Thiswork’suseordiscussion
ofMATLABⓇsoftwareorrelatedproductsdoesnotconstituteendorsementorsponsorshipbyThe
MathWorksofaparticularpedagogicalapproachorparticularuseoftheMATLABⓇsoftware.
Whilethepublisherandauthorshaveusedtheirbesteffortsinpreparingthiswork,theymakeno
representationsorwarrantieswithrespecttotheaccuracyorcompletenessofthecontentsofthisworkand
specificallydisclaimallwarranties,includingwithoutlimitationanyimpliedwarrantiesofmerchantability
orfitnessforaparticularpurpose.Nowarrantymaybecreatedorextendedbysalesrepresentatives,written
salesmaterialsorpromotionalstatementsforthiswork.Thefactthatanorganization,website,orproductis
referredtointhisworkasacitationand/orpotentialsourceoffurtherinformationdoesnotmeanthatthe
publisherandauthorsendorsetheinformationorservicestheorganization,website,orproductmay
provideorrecommendationsitmaymake.Thisworkissoldwiththeunderstandingthatthepublisherisnot
engagedinrenderingprofessionalservices.Theadviceandstrategiescontainedhereinmaynotbesuitable
foryoursituation.Youshouldconsultwithaspecialistwhereappropriate.Further,readersshouldbeaware
thatwebsiteslistedinthisworkmayhavechangedordisappearedbetweenwhenthisworkwaswrittenand
whenitisread.Neitherthepublishernorauthorsshallbeliableforanylossofprofitoranyother
commercialdamages,includingbutnotlimitedtospecial,incidental,consequential,orotherdamages.
LibraryofCongressCataloging-in-Publicationdata
Names:Rojo-Álvarez,JoséLuis,–author.|Martínez-Ramón,Manel,–author.|
Muñoz-Marí,Jordi,author.|Camps-Valls,Gustau,–author.
Title:Digitalsignalprocessingwithkernelmethods/byDr.JoséLuisRojo-Álvarez,
Dr.ManelMartínez-Ramón,Dr.JordiMuñoz-Marí,Dr.GustauCamps-Valls.
Description:Firstedition.|Hoboken,NJ:JohnWiley&Sons,.|Includesbibliographical
referencesandindex.|
Identifiers:LCCN(print)|LCCN(ebook)|ISBN(pdf)|
ISBN(epub)|ISBN(cloth)
Subjects:LCSH:Signalprocessing–Digitaltechniques.
Classification:LCCTK.(ebook)|LCCTK..R(print)|DDC./–dc
LCrecordavailableathttps://lccn.loc.gov/
CoverdesignbyWiley
Coverimage:©AF-studio/Gettyimages
Setin/ptWarnockbySPiGlobal,Pondicherry,India
v
Contents
AbouttheAuthors xiii
Preface xvii
Acknowledgements xxi
ListofAbbreviations xxiii
PartI FundamentalsandBasicElements
1 FromSignalProcessingtoMachineLearning
. ANewScienceisBorn:SignalProcessing
.. SignalProcessingBeforeBeingCoined
.. :BirthoftheInformationAge
.. s:AudioEngineeringCatalyzesSignalProcessing
. FromAnalogtoDigitalSignalProcessing
.. s:DigitalSignalProcessingBegins
.. s:DigitalSignalProcessingBecomesPopular
.. s:SiliconMeetsDigitalSignalProcessing
. DigitalSignalProcessingMeetsMachineLearning
.. s:NewApplicationAreas
.. s:NeuralNetworks,FuzzyLogic,andGeneticOptimization
. RecentMachineLearninginDigitalSignalProcessing
.. TraditionalSignalAssumptionsAreNoLongerValid
.. EncodingPriorKnowledge
.. LearningandKnowledgefromData
.. FromMachineLearningtoDigitalSignalProcessing
.. FromDigitalSignalProcessingtoMachineLearning
2 IntroductiontoDigitalSignalProcessing
. OutlineoftheSignalProcessingField
.. FundamentalsonSignalsandSystems
.. DigitalFiltering
.. SpectralAnalysis
.. Deconvolution
.. Interpolation
vi Contents
.. SystemIdentification
.. BlindSourceSeparation
. FromTime–FrequencytoCompressedSensing
.. Time–FrequencyDistributions
.. WaveletTransforms
.. Sparsity,CompressedSensing,andDictionaryLearning
. MultidimensionalSignalsandSystems
.. MultidimensionalSignals
.. MultidimensionalSystems
. SpectralAnalysisonManifolds
.. TheoreticalFundamentals
.. LaplacianMatrices
. TutorialsandApplicationExamples
.. RealandComplexSignalProcessingandRepresentations
.. Convolution,FourierTransform,andSpectrum
.. Continuous-TimeSignalsandSystems
.. FilteringCardiacSignals
.. NonparametricSpectrumEstimation
.. ParametricSpectrumEstimation
.. SourceSeparation
.. Time–FrequencyRepresentationsandWavelets
.. ExamplesforSpectralAnalysisonManifolds
. QuestionsandProblems
3 SignalProcessingModels
. Introduction
. VectorSpaces,Basis,andSignalModels
.. BasicOperationsforVectors
.. VectorSpaces
.. HilbertSpaces
.. SignalModels
.. ComplexSignalModels
.. StandardNoiseModelsinDSP
.. TheRoleoftheCostFunction
.. TheRoleoftheRegularizer
. DigitalSignalProcessingModels
.. SinusoidalSignalModels
.. SystemIdentificationSignalModels
.. SincInterpolationModels
.. SparseDeconvolution
.. ArrayProcessing
. TutorialsandApplicationExamples
.. ExamplesofNoiseModels
.. AutoregressiveExogenousSystemIdentificationModels
.. NonlinearSystemIdentificationUsingVolterraModels
.. SinusoidalSignalModels
Contents vii
.. Sinc-basedInterpolation
.. SparseDeconvolution
.. ArrayProcessing
. QuestionsandProblems
.A MATLABsimpleInterpToolboxStructure
4 KernelFunctionsandReproducingKernelHilbertSpaces
. Introduction
. KernelFunctionsandMappings
.. MeasuringSimilaritywithKernels
.. Positive-DefiniteKernels
.. ReproducingKernelinHilbertSpaceandReproducingProperty
.. Mercer’sTheorem
. KernelProperties
.. Tikhonov’sRegularization
.. RepresenterTheoremandRegularizationProperties
.. BasicOperationswithKernels
. ConstructingKernelFunctions
.. StandardKernels
.. PropertiesofKernels
.. EngineeringSignalProcessingKernels
. ComplexReproducingKernelinHilbertSpaces
. SupportVectorMachineElementsforRegressionandEstimation
.. SupportVectorRegressionSignalModelandCostFunction
.. MinimizingFunctional
. TutorialsandApplicationExamples
.. KernelCalculationsandKernelMatrices
.. BasicOperationswithKernels
.. ConstructingKernels
.. ComplexKernels
.. ApplicationExampleforSupportVectorRegressionElements
. ConcludingRemarks
. QuestionsandProblems
PartII FunctionApproximationandAdaptiveFiltering
5 ASupportVectorMachineSignalEstimationFramework
. Introduction
. AFrameworkforSupportVectorMachineSignalEstimation
. PrimalSignalModelsforSupportVectorMachineSignalProcessing
.. NonparametricSpectrumandSystemIdentification
.. OrthogonalFrequencyDivisionMultiplexingDigitalCommunications
.. ConvolutionalSignalModels
.. ArrayProcessing
. TutorialsandApplicationExamples
viii Contents
.. NonparametricSpectralAnalysiswithPrimalSignalModels
.. SystemIdentificationwithPrimalSignalModelγ-filter
.. ParametricSpectralDensityEstimationwithPrimalSignalModels
.. TemporalReferenceArrayProcessingwithPrimalSignalModels
.. SincInterpolationwithPrimalSignalModels
.. OrthogonalFrequencyDivisionMultiplexingwithPrimalSignalModels
. QuestionsandProblems
6 ReproducingKernelHilbertSpaceModelsforSignalProcessing
. Introduction
. ReproducingKernelHilbertSpaceSignalModels
.. KernelAutoregressiveExogenousIdentification
.. KernelFiniteImpulseResponseandtheγ-filter
.. KernelArrayProcessingwithSpatialReference
.. KernelSemiparametricRegression
. TutorialsandApplicationExamples
.. NonlinearSystemIdentificationwithSupportVector
Machine–AutoregressiveandMovingAverage
.. NonlinearSystemIdentificationwiththeγ-filter
.. ElectricNetworkModelingwithSemiparametricRegression
.. PromotionalData
.. SpatialandTemporalAntennaArrayKernelProcessing
. QuestionsandProblems
7 DualSignalModelsforSignalProcessing
. Introduction
. DualSignalModelElements
. DualSignalModelInstantiations
.. DualSignalModelforNonuniformSignalInterpolation
.. DualSignalModelforSparseSignalDeconvolution
.. SpectrallyAdaptedMercerKernels
. TutorialsandApplicationExamples
.. NonuniformInterpolationwiththeDualSignalModel
.. SparseDeconvolutionwiththeDualSignalModel
.. DopplerUltrasoundProcessingforFaultDetection
.. SpectrallyAdaptedMercerKernels
.. InterpolationofHeartRateVariabilitySignals
.. DenoisinginCardiacMotion-ModeDopplerUltrasoundImages
.. IndoorLocationfromMobileDevicesMeasurements
.. ElectroanatomicalMapsinCardiacNavigationSystems
. QuestionsandProblems
8 AdvancesinKernelRegressionandFunctionApproximation
. Introduction
. Kernel-BasedRegressionMethods
.. AdvancesinSupportVectorRegression
.. Multi-outputSupportVectorRegression
Contents ix
.. KernelRidgeRegression
.. KernelSignal-to-NoiseRegression
.. Semi-supervisedSupportVectorRegression
.. ModelSelectioninKernelRegressionMethods
. BayesianNonparametricKernelRegressionModels
.. GaussianProcessRegression
.. RelevanceVectorMachines
. TutorialsandApplicationExamples
.. ComparingSupportVectorRegression,RelevanceVectorMachines,and
GaussianProcessRegression
.. Profile-DependentSupportVectorRegression
.. Multi-outputSupportVectorRegression
.. KernelSignal-to-NoiseRatioRegression
.. Semi-supervisedSupportVectorRegression
.. BayesianNonparametricModel
.. GaussianProcessRegression
.. RelevanceVectorMachines
. ConcludingRemarks
. QuestionsandProblems
9 AdaptiveKernelLearningforSignalProcessing
. Introduction
. LinearAdaptiveFiltering
.. LeastMeanSquaresAlgorithm
.. RecursiveLeast-SquaresAlgorithm
. KernelAdaptiveFiltering
. KernelLeastMeanSquares
.. DerivationofKernelLeastMeanSquares
.. ImplementationChallengesandDualFormulation
.. ExampleonPredictionoftheMackey–GlassTimeSeries
.. PracticalKernelLeastMeanSquaresAlgorithms
. KernelRecursiveLeastSquares
.. KernelRidgeRegression
.. DerivationofKernelRecursiveLeastSquares
.. PredictionoftheMackey–GlassTimeSerieswithKernelRecursiveLeast
Squares
.. BeyondtheStationaryModel
.. ExampleonNonlinearChannelIdentificationandReconvergence
. ExplicitRecursivityforAdaptiveKernelModels
.. RecursivityinHilbertSpaces
.. RecursiveFiltersinReproducingKernelHilbertSpaces
. OnlineSparsificationwithKernels
.. SparsitybyConstruction
.. SparsitybyPruning
. ProbabilisticApproachestoKernelAdaptiveFiltering
.. GaussianProcessesandKernelRidgeRegression
.. OnlineRecursiveSolutionforGaussianProcessesRegression
x Contents
.. KernelRecursiveLeastSquaresTracker
.. ProbabilisticKernelLeastMeanSquares
. FurtherReading
.. SelectionofKernelParameters
.. Multi-KernelAdaptiveFiltering
.. RecursiveFilteringinKernelHilbertSpaces
. TutorialsandApplicationExamples
.. KernelAdaptiveFilteringToolbox
.. PredictionofaRespiratoryMotionTimeSeries
.. OnlineRegressionontheKINKDataset
.. TheMackey–GlassTimeSeries
.. ExplicitRecursivityonReproducingKernelinHilbertSpace
andElectroencephalogramPrediction
.. AdaptiveAntennaArrayProcessing
. QuestionsandProblems
PartIII Classification,Detection,andFeatureExtraction
10 SupportVectorMachineandKernelClassificationAlgorithms
. Introduction
. SupportVectorMachineandKernelClassifiers
.. SupportVectorMachines
.. MulticlassandMultilabelSupportVectorMachines
.. Least-SquaresSupportVectorMachine
.. KernelFisher’sDiscriminantAnalysis
. AdvancesinKernel-BasedClassification
.. LargeMarginFiltering
.. Semi-supervisedLearning
.. MultipleKernelLearning
.. Structured-OutputLearning
.. ActiveLearning
. Large-ScaleSupportVectorMachines
.. Large-ScaleSupportVectorMachineImplementations
.. RandomFourierFeatures
.. ParallelSupportVectorMachine
.. Outlook
. TutorialsandApplicationExamples
.. ExamplesofSupportVectorMachineClassification
.. ExampleofLeast-SquaresSupportVectorMachine
.. Kernel-FilteringSupportVectorMachineforBrain–ComputerInterface
SignalClassification
.. ExampleofLaplacianSupportVectorMachine
.. ExampleofGraph-BasedLabelPropagation
.. ExamplesofMultipleKernelLearning
. ConcludingRemarks
. QuestionsandProblems
Description:A realistic and comprehensive review of joint approaches to machine learning and signal processing algorithms, with application to communications, multimedia, and biomedical engineering systems Digital Signal Processing with Kernel Methods reviews the milestones in the mixing of classical digital si