金融经典，好东西，物有所值。

NeuralNetworks
inFinance:
Gaining
PredictiveEdge
intheMarket
PaulD.McNelis

Contents
Prefacexi
1Introduction1
1.1Forecasting,Classi?cation,andDimensionality
Reduction ..........................1
1.2Synergies ...........................4
1.3TheInterfaceProblems ...................6
1.4PlanoftheBook ......................8
IEconometricFoundations11
2WhatAreNeuralNetworks?13
2.1LinearRegressionModel ..................13
2.2GARCHNonlinearModels .................15
2.2.1PolynomialApproximation .............17
2.2.2OrthogonalPolynomials ...............18
2.3ModelTypology .......................20
2.4WhatIsANeuralNetwork? ................21
2.4.1FeedforwardNetworks ................21
2.4.2SquasherFunctions .................24
2.4.3RadialBasisFunctions ...............28
2.4.4RidgeletNetworks ..................29
2.4.5JumpConnections ..................30
2.4.6MultilayeredFeedforwardNetworks ........32

2.4.7RecurrentNetworks .................34
2.4.8NetworkswithMultipleOutputs ..........36
2.5NeuralNetworkSmooth-TransitionRegimeSwitching
Models ............................38
2.5.1Smooth-TransitionRegimeSwitchingModels...38
2.5.2NeuralNetworkExtensions .............39
2.6NonlinearPrincipalComponents:Intrinsic
Dimensionality ........................41
2.6.1LinearPrincipalComponents ............42
2.6.2NonlinearPrincipalComponents ..........44
2.6.3ApplicationtoAssetPricing ............46
2.7NeuralNetworksandDiscreteChoice ...........49
2.7.1DiscriminantAnalysis ................49
2.7.2LogitRegression ...................50
2.7.3ProbitRegression ..................51
2.7.4WeibullRegression .................52
2.7.5NeuralNetworkModelsforDiscreteChoice ....52
2.7.6ModelswithMultinomialOrderedChoice .....53
2.8TheBlackBoxCriticismandDataMining ........55
2.9Conclusion ..........................57
2.9.1MATLABProgramNotes ..............58
2.9.2SuggestedExercises .................58
3EstimationofaNetworkwithEvolutionaryComputation59
3.1DataPreprocessing .....................59
3.1.1Stationarity:Dickey-FullerTest ...........59
3.1.2SeasonalAdjustment:CorrectionforCalendar
E?ects ........................61
3.1.3DataScaling .....................64
3.2TheNonlinearEstimationProblem ............65
3.2.1LocalGradient-BasedSearch:TheQuasi-Newton
MethodandBackpropagation ...........67
3.2.2StochasticSearch:SimulatedAnnealing ......70
3.2.3EvolutionaryStochasticSearch:TheGenetic
Algorithm ......................72
3.2.4EvolutionaryGeneticAlgorithms ..........75
3.2.5Hybridization:CouplingGradient-Descent,
Stochastic,andGeneticSearchMethods ......75
3.3RepeatedEstimationandThickModels ..........77
3.4MATLABExamples:NumericalOptimizationand
NetworkPerformance ....................78
3.4.1NumericalOptimization ...............78
3.4.2ApproximationwithPolynomialsand
NeuralNetworks ...................80

Contentsvii
3.5Conclusion ..........................83
3.5.1MATLABProgramNotes ..............83
3.5.2SuggestedExercises .................84
4EvaluationofNetworkEstimation85
4.1In-SampleCriteria ......................85
4.1.1GoodnessofFitMeasure ..............86
4.1.2Hannan-QuinnInformationCriterion .......86
4.1.3SerialIndependence:Ljung-BoxandMcLeod-Li
Tests .........................86
4.1.4Symmetry ......................89
4.1.5Normality ......................89
4.1.6NeuralNetworkTestforNeglectedNonlinearity:
Lee-White-GrangerTest ..............90
4.1.7Brock-Deckert-ScheinkmanTestforNonlinear
Patterns .......................91
4.1.8SummaryofIn-SampleCriteria ...........93
4.1.9MATLABExample .................93
4.2Out-of-SampleCriteria ...................94
4.2.1RecursiveMethodology ...............95
4.2.2RootMeanSquaredErrorStatistic .........96
4.2.3Diebold-MarianoTestforOut-of-SampleErrors..96
4.2.4Harvey,Leybourne,andNewboldSizeCorrection
ofDiebold-MarianoTest ..............97
4.2.5Out-of-SampleComparisonwithNestedModels..98
4.2.6SuccessRatioforSignPredictions:Directional
Accuracy .......................99
4.2.7PredictiveStochasticComplexity ..........100
4.2.8Cross-Validationandthe.632Bootstrapping
Method ........................101
4.2.9DataRequirements:HowLargeforPredictive
Accuracy? ......................102
4.3InterpretiveCriteriaandSigni?canceofResults ......104
4.3.1AnalyticDerivatives .................105
4.3.2FiniteDi?erences ..................106
4.3.3DoesItMatter? ...................107
4.3.4MATLABExample:AnalyticandFinite
Di?erences ......................107
4.3.5BootstrappingforAssessingSigni?cance ......108
4.4ImplementationStrategy ..................109
4.5Conclusion ..........................110
4.5.1MATLABProgramNotes ..............110
4.5.2SuggestedExercises .................111

viiiContents
IIApplicationsandExamples113
5EstimatingandForecastingwithArti?cialData115
5.1Introduction .........................115
5.2StochasticChaosModel ...................117
5.2.1In-SamplePerformance ...............118
5.2.2Out-of-SamplePerformance .............120
5.3StochasticVolatility/JumpDi?usionModel ........122
5.3.1In-SamplePerformance ...............123
5.3.2Out-of-SamplePerformance .............125
5.4TheMarkovRegimeSwitchingModel ...........125
5.4.1In-SamplePerformance ...............128
5.4.2Out-of-SamplePerformance .............130
5.5VolatalityRegimeSwitchingModel ............130
5.5.1In-SamplePerformance ...............132
5.5.2Out-of-SamplePerformance .............132
5.6DistortedLong-MemoryModel ...............135
5.6.1In-SamplePerformance ...............136
5.6.2Out-of-SamplePerformance .............137
5.7Black-SholesOptionPricingModel:ImpliedVolatility
Forecasting ..........................137
5.7.1In-SamplePerformance ...............140
5.7.2Out-of-SamplePerformance .............142
5.8Conclusion ..........................142
5.8.1MATLABProgramNotes ..............142
5.8.2SuggestedExercises .................143
6TimesSeries:ExamplesfromIndustryandFinance145
6.1ForecastingProductionintheAutomotiveIndustry...145
6.1.1TheData .......................146
6.1.2ModelsofQuantityAdjustment ..........148
6.1.3In-SamplePerformance ...............150
6.1.4Out-of-SamplePerformance .............151
6.1.5InterpretationofResults ..............152
6.2CorporateBonds:WhichFactorsDeterminethe
Spreads? ...........................156
6.2.1TheData .......................157
6.2.2AModelfortheAdjustmentofSpreads ......157
6.2.3In-SamplePerformance ...............160
6.2.4Out-of-SamplePerformance .............160
6.2.5InterpretationofResults ..............161

Contentsix
6.3Conclusion ..........................165
6.3.1MATLABProgramNotes ..............166
6.3.2SuggestedExercises .................166
7In?ationandDe?ation:HongKongandJapan167
7.1HongKong ..........................168
7.1.1TheData .......................169
7.1.2ModelSpeci?cation .................174
7.1.3In-SamplePerformance ...............177
7.1.4Out-of-SamplePerformance .............177
7.1.5InterpretationofResults ..............178
7.2Japan ............................182
7.2.1TheData .......................184
7.2.2ModelSpeci?cation .................189
7.2.3In-SamplePerformance ...............189
7.2.4Out-of-SamplePerformance .............190
7.2.5InterpretationofResults ..............191
7.3Conclusion ..........................196
7.3.1MATLABProgramNotes ..............196
7.3.2SuggestedExercises .................196
8Classi?cation:CreditCardDefaultandBankFailures199
8.1CreditCardRisk ......................200
8.1.1TheData .......................200
8.1.2In-SamplePerformance ...............200
8.1.3Out-of-SamplePerformance .............202
8.1.4InterpretationofResults ..............203
8.2BankingIntervention ....................204
8.2.1TheData .......................204
8.2.2In-SamplePerformance ...............205
8.2.3Out-of-SamplePerformance .............207
8.2.4InterpretationofResults ..............208
8.3Conclusion ..........................209
8.3.1MATLABProgramNotes ..............210
8.3.2SuggestedExercises .................210
9DimensionalityReductionandImpliedVolatility
Forecasting211
9.1HongKong ..........................212
9.1.1TheData .......................212
9.1.2In-SamplePerformance ...............213
9.1.3Out-of-SamplePerformance .............214

xContents
9.2UnitedStates ........................216
9.2.1TheData .......................216
9.2.2In-SamplePerformance ...............216
9.2.3Out-of-SamplePerformance .............218
9.3Conclusion ..........................219
9.3.1MATLABProgramNotes ..............220
9.3.2SuggestedExercises .................220
Bibliography221
Index233

[此贴子已经被作者于2008-5-1 8:09:08编辑过]