Elsevier S & T

Neural Networks in Finance: Gaining Predictive Edge in the Market

McNelis, Paul D.

In stock
No reviews
38.500 KD inc. VAT
License
Table of contents
  • Cover
  • Table of contentsv
  • 1 Introduction1
  • 1.1 Forecasting, Classification, and Dimensionality Reduction1
  • 1.2 Synergies4
  • 1.3 The Interface Problems6
  • 1.4 Plan of the Book8
  • Part I Econometric Foundations11
  • 2 What Are Neural Networks?13
  • 2.1 Linear Regression Model13
  • 2.2 GARCH Nonlinear Models15
  • 2.2.1 Polynomial Approximation17
  • 2.2.2 Orthogonal Polynomials18
  • 2.3 Model Typology20
  • 2.4 What Is A Neural Network?21
  • 2.4.1 Feedforward Networks21
  • 2.4.2 Squasher Functions24
  • 2.4.3 Radial Basis Functions28
  • 2.4.4 Ridgelet Networks29
  • 2.4.5 Jump Connections30
  • 2.4.6 Multilayered Feedforward Networks32
  • 2.4.7 Recurrent Networks34
  • 2.4.8 Networks with Multiple Outputs36
  • 2.5 Neural Network Smooth-Transition Regime Switching Models38
  • 2.5.1 Smooth-Transition Regime Switching Models38
  • 2.5.2 Neural Network Extensions39
  • 2.6 Nonlinear Principal Components: Intrinsic Dimensionality41
  • 2.6.1 Linear Principal Components42
  • 2.6.2 Nonlinear Principal Components44
  • 2.6.3 Application to Asset Pricing46
  • 2.7 Neural Networks and Discrete Choice49
  • 2.7.1 Discriminant Analysis49
  • 2.7.2 Logit Regression50
  • 2.7.3 Probit Regression51
  • 2.7.4 Weibull Regression52
  • 2.7.5 Neural Network Models for Discrete Choice52
  • 2.7.6 Models with Multinomial Ordered Choice53
  • 2.8 The Black Box Criticism and Data Mining55
  • 2.9 Conclusion57
  • 2.9.1 MATLAB Program Notes58
  • 2.9.2 Suggested Exercises58
  • 3 Estimation of a Network with Evolutionary Computation59
  • 3.1 Data Preprocessing59
  • 3.1.1 Stationarity: Dickey-Fuller Test59
  • 3.1.2 Seasonal Adjustment: Correction for Calendar Effects61
  • 3.1.3 Data Scaling64
  • 3.2 The Nonlinear Estimation Problem65
  • 3.2.1 Local Gradient-Based Search: The Quasi-Newton Method and Backpropagation67
  • 3.2.2 Stochastic Search: Simulated Annealing70
  • 3.2.3 Evolutionary Stochastic Search: The Genetic Algorithm72
  • 3.2.4 Evolutionary Genetic Algorithms75
  • 3.2.5 Hybridization: Coupling Gradient-Descent, Stochastic, and Genetic Search Methods75
  • 3.3 Repeated Estimation and Thick Models77
  • 3.4 MATLAB Examples: Numerical Optimization and Network Performance78
  • 3.4.1 Numerical Optimization78
  • 3.4.2 Approximation with Polynomials and Neural Networks80
  • 3.5 Conclusion83
  • 3.5.1 MATLAB Program Notes83
  • 3.5.2 Suggested Exercises84
  • 4 Evaluation of Network Estimation85
  • 4.1 In-Sample Criteria85
  • 4.1.1 Goodness of Fit Measure86
  • 4.1.2 Hannan-Quinn Information Criterion86
  • 4.1.3 Serial Independence: Ljung-Box and McLeod-Li Tests86
  • 4.1.5 Normality89
  • 4.1.6 Neural Network Test for Neglected Nonlinearity: Lee-White-Granger Test90
  • 4.1.7 Brock-Deckert-Scheinkman Test for Nonlinear Patterns91
  • 4.1.8 Summary of In-Sample Criteria93
  • 4.1.9 MATLAB Example93
  • 4.2 Out-of-Sample Criteria94
  • 4.2.1 Recursive Methodology95
  • 4.2.2 Root Mean Squared Error Statistic96
  • 4.2.3 Diebold-Mariano Test for Out-of-Sample Errors96
  • 4.2.4 Harvey, Leybourne, and Newbold Size Correction of Diebold-Mariano Test97
  • 4.2.5 Out-of-Sample Comparison with Nested Models98
  • 4.2.6 Success Ratio for Sign Predictions: Directional Accuracy99
  • 4.2.7 Predictive Stochastic Complexity100
  • 4.2.8 Cross-Validation and the .632 Bootstrapping Method101
  • 4.2.9 Data Requirements: How Large for Predictive Accuracy?102
  • 4.3 Interpretive Criteria and Significance of Results104
  • 4.3.1 Analytic Derivatives105
  • 4.3.2 Finite Differences106
  • 4.3.3 Does It Matter?107
  • 4.3.4 MATLAB Example: Analytic and Finite Differences107
  • 4.3.5 Bootstrapping for Assessing Significance108
  • 4.4 Implementation Strategy109
  • 4.5 Conclusion110
  • 4.5.1 MATLAB Program Notes110
  • 4.5.2 Suggested Exercises111
  • Part II Applications and Examples113
  • 5 Estimating and Forecasting with Artificial Data115
  • 5.1 Introduction115
  • 5.2 Stochastic Chaos Model117
  • 5.2.1 In-Sample Performance118
  • 5.2.2 Out-of-Sample Performance120
  • 5.3 Stochastic Volatility/Jump Diffusion Model122
  • 5.3.1 In-Sample Performance123
  • 5.3.2 Out-of-Sample Performance125
  • 5.4 The Markov Regime Switching Model125
  • 5.4.1 In-Sample Performance128
  • 5.4.2 Out-of-Sample Performance130
  • 5.5 Volatility Regime Switching Model130
  • 5.5.1 In-Sample Performance132
  • 5.5.2 Out-of-Sample Performance132
  • 5.6 Distorted Long-Memory Model135
  • 5.6.1 In-Sample Performance136
  • 5.6.2 Out-of-Sample Performance137
  • 5.7 Black-Sholes Option Pricing Model: Implied Volatility Forecasting137
  • 5.7.1 In-Sample Performance140
  • 5.7.2 Out-of-Sample Performance142
  • 5.8 Conclusion142
  • 5.8.1 MATLAB Program Notes142
  • 5.8.2 Suggested Exercises143
  • 6 Times Series: Examples from Industry and Finance145
  • 6.1 Forecasting Production in the Automotive Industry145
  • 6.1.1 The Data146
  • 6.1.2 Models of Quantity Adjustment148
  • 6.1.3 In-Sample Performance150
  • 6.1.4 Out-of-Sample Performance151
  • 6.1.5 Interpretation of Results152
  • 6.2 Corporate Bonds: Which Factors Determine the Spreads?156
  • 6.2.1 The Data157
  • 6.2.2 A Model for the Adjustment of Spreads157
  • 6.2.3 In-Sample Performance160
  • 6.2.4 Out-of-Sample Performance160
  • 6.2.5 Interpretation of Results161
  • 6.3 Conclusion165
  • 6.3.1 MATLAB Program Notes166
  • 6.3.2 Suggested Exercises166
  • 7 Inflation and Deflation: Hong Kong and Japan167
  • 7.1 Hong Kong168
  • 7.1.1 The Data169
  • 7.1.2 Model Specification174
  • 7.1.3 In-Sample Performance177
  • 7.1.4 Out-of-Sample Performance177
  • 7.1.5 Interpretation of Results178
  • 7.2 Japan182
  • 7.2.1 The Data184
  • 7.2.2 Model Specification189
  • 7.2.3 In-Sample Performance189
  • 7.2.4 Out-of-Sample Performance190
  • 7.2.5 Interpretation of Results191
  • 7.3 Conclusion196
  • 7.3.1 MATLAB Program Notes196
  • 7.3.2 Suggested Exercises196
  • 8 Classification: Credit Card Default and Bank Failures199
  • 8.1 Credit Card Risk200
  • 8.1.1 The Data200
  • 8.1.2 In-Sample Performance200
  • 8.1.3 Out-of-Sample Performance202
  • 8.1.4 Interpretation of Results203
  • 8.2 Banking Intervention204
  • 8.2.1 The Data204
  • 8.2.2 In-Sample Performance205
  • 8.2.3 Out-of-Sample Performance207
  • 8.2.4 Interpretation of Results208
  • 8.3 Conclusion209
  • 8.3.1 MATLAB Program Notes210
  • 8.3.2 Suggested Exercises210
  • 9 Dimensionality Reduction and Implied Volatility Forecasting211
  • 9.1 Hong Kong212
  • 9.1.1 The Data212
  • 9.1.2 In-Sample Performance213
  • 9.1.3 Out-of-Sample Performance214
  • 9.2 United States216
  • 9.2.1 The Data216
  • 9.2.2 In-Sample Performance216
  • 9.2.3 Out-of-Sample Performance218
  • 9.3 Conclusion219
  • 9.3.1 MATLAB Program Notes220
  • 9.3.2 Suggested Exercises220
  • Bibliography221
  • Index233
Book details
  • Vendor Elsevier S & T
  • SKU 9780124859678
  • ISBN-13 9780080479651
  • Author McNelis, Paul D.
  • Category Business & Economics
  • Subject Finance

Do you have questions about this book?

Ask an expert!

Description eBook delivery & access Reviews
This book explores the intuitive appeal of neural networks and the genetic algorithm in finance. It demonstrates how neural networks used in combination with evolutionary computation outperform classical econometric methods for accuracy in forecasting, classification and dimensionality reduction.

McNelis utilizes a variety of examples, from forecasting automobile production and corporate bond spread, to inflation and deflation processes in Hong Kong and Japan, to credit card default in Germany to bank failures in Texas, to cap-floor volatilities in New York and Hong Kong.

* Offers a balanced, critical review of the neural network methods and genetic algorithms used in finance
* Includes numerous examples and applications
* Numerical illustrations use MATLAB code and the book is accompanied by a website