Nonlinear Time Series Analysis
Description
More details
Other editions
Additional editions
Persons
RUEY S. TSAY, PHD, is H.G.B. Alexander Professor of Econometrics and Statistics at The University of Chicago Booth School of Business. He is a fellow of the American Statistical Association and the Institute of Mathematical Statistics.Dr. Tsay is author of Analysis of Financial Time Series, Multivariate Time Series Analysis, and An Introduction to Analysis of Financial Data with R all published by Wiley.
RONG CHEN, PHD, is Distinguished Professor of Statistics and Director of the Master programs in Financial Statistics and Risk Management and in Data Science at Rutgers University. He is a fellow of the American Statistical Association and the Institute of Mathematical Statistics.
Content
Preface xiii
1 Why Should We Care About Nonlinearity? 1
1.1 Some Basic Concepts 2
1.2 Linear Time Series 3
1.3 Examples of Nonlinear Time Series 3
1.4 Nonlinearity Tests 20
1.4.1 Nonparametric Tests 21
1.4.2 Parametric Tests 31
1.5 Exercises 38
References 39
2 Univariate Parametric Nonlinear Models 41
2.1 A General Formulation 41
2.1.1 Probability Structure 42
2.2 Threshold Autoregressive Models 43
2.2.1 A Two-regime TAR Model 44
2.2.2 Properties of Two-regime TAR(1) Models 45
2.2.3 Multiple-regime TAR Models 48
2.2.4 Estimation of TAR Models 50
2.2.5 TAR Modeling 52
2.2.6 Examples 55
2.2.7 Predictions of TAR Models 62
2.3 Markov Switching Models 63
2.3.1 Properties of Markov Switching Models 66
2.3.2 Statistical Inference of the State Variable 66
2.3.3 Estimation of Markov Switching Models 69
2.3.4 Selecting the Number of States 75
2.3.5 Prediction of Markov Switching Models 75
2.3.6 Examples 76
2.4 Smooth Transition Autoregressive Models 92
2.5 Time-varying Coefficient Models 99
2.5.1 Functional Coefficient AR Models 99
2.5.2 Time-varying Coefficient AR Models 104
2.6 Appendix: Markov Chains 111
2.7 Exercises 114
References 116
3 Univariate Nonparametric Models 119
3.1 Kernel Smoothing 119
3.2 Local Conditional Mean 125
3.3 Local Polynomial Fitting 129
3.4 Splines 134
3.4.1 Cubic and B-Splines 138
3.4.2 Smoothing Splines 141
3.5 Wavelet Smoothing 145
3.5.1 Wavelets 145
3.5.2 The Wavelet Transform 147
3.5.3 Thresholding and Smoothing 150
3.6 Nonlinear Additive Models 158
3.7 Index Model and Sliced Inverse Regression 164
3.8 Exercises 169
References 170
4 Neural Networks, Deep Learning, and Tree-based Methods 173
4.1 Neural Networks 173
4.1.1 Estimation or Training of Neural Networks 176
4.1.2 An Example 179
4.2 Deep Learning 181
4.2.1 Deep Belief Nets 182
4.2.2 Demonstration 184
4.3 Tree-based Methods 195
4.3.1 Decision Trees 195
4.3.2 Random Forests 212
4.4 Exercises 214
References 215
5 Analysis of Non-Gaussian Time Series 217
5.1 Generalized Linear Time Series Models 218
5.1.1 Count Data and GLARMA Models 220
5.2 Autoregressive Conditional Mean Models 229
5.3 Martingalized GARMA Models 232
5.4 Volatility Models 234
5.5 Functional Time Series 245
5.5.1 Convolution FAR models 248
5.5.2 Estimation of CFAR Models 251
5.5.3 Fitted Values and Approximate Residuals 253
5.5.4 Prediction 253
5.5.5 Asymptotic Properties 254
5.5.6 Application 254
Appendix: Discrete Distributions for Count Data 260
5.6 Exercises 261
References 263
6 State Space Models 265
6.1 A General Model and Statistical Inference 266
6.2 Selected Examples 269
6.2.1 Linear Time Series Models 269
6.2.2 Time Series with Observational Noises 271
6.2.3 Time-varying Coefficient Models 272
6.2.4 Target Tracking 273
6.2.5 Signal Processing in Communications 279
6.2.6 Dynamic Factor Models 283
6.2.7 Functional and Distributional Time Series 284
6.2.8 Markov Regime Switching Models 289
6.2.9 Stochastic Volatility Models 290
6.2.10 Non-Gaussian Time Series 291
6.2.11 Mixed Frequency Models 291
6.2.12 Other Applications 292
6.3 Linear Gaussian State Space Models 293
6.3.1 Filtering and the Kalman Filter 293
6.3.2 Evaluating the likelihood function 295
6.3.3 Smoothing 297
6.3.4 Prediction and Missing Data 299
6.3.5 Sequential Processing 300
6.3.6 Examples and R Demonstrations 300
6.4 Exercises 325
References 327
7 Nonlinear State Space Models 335
7.1 Linear and Gaussian Approximations 335
7.1.1 Kalman Filter for Linear Non-Gaussian Systems 336
7.1.2 Extended Kalman Filters for Nonlinear Systems 336
7.1.3 Gaussian Sum Filters 338
7.1.4 The Unscented Kalman Filter 339
7.1.5 Ensemble Kalman Filters 341
7.1.6 Examples and R implementations 342
7.2 Hidden Markov Models 351
7.2.1 Filtering 351
7.2.2 Smoothing 352
7.2.3 The Most Likely State Path: the Viterbi Algorithm 355
7.2.4 Parameter Estimation: the Baum-Welch Algorithm 356
7.2.5 HMM Examples and R Implementation 358
7.3 Exercises 371
References 372
8 Sequential Monte Carlo 375
8.1 A Brief Overview of Monte Carlo Methods 376
8.1.1 General Methods of Generating Random Samples 378
8.1.2 Variance Reduction Methods 384
8.1.3 Importance Sampling 387
8.1.4 Markov Chain Monte Carlo 398
8.2 The SMC Framework 402
8.3 Design Issue I: Propagation 410
8.3.1 Proposal Distributions 411
8.3.2 Delay Strategy (Lookahead) 415
8.4 Design Issue II: Resampling 421
8.4.1 The Priority Score 422
8.4.2 Choice of Sampling Methods in Resampling 423
8.4.3 Resampling Schedule 425
8.4.4 Benefits of Resampling 426
8.5 Design Issue III: Inference 428
8.6 Design Issue IV: Marginalization and the Mixture Kalman Filter 429
8.6.1 Conditional Dynamic Linear Models 429
8.6.2 Mixture Kalman Filters 430
8.7 Smoothing with SMC 433
8.7.1 Simple Weighting Approach 433
8.7.2 Weight Marginalization Approach 434
8.7.3 Two-filter Sampling 436
8.8 Parameter Estimation with SMC 438
8.8.1 Maximum Likelihood Estimation 438
8.8.2 Bayesian Parameter Estimation 441
8.8.3 Varying Parameter Approach 441
8.9 Implementation Considerations 442
8.10 Examples and R Implementation 444
8.10.1 R Implementation of SMC: Generic SMC and Resampling Methods 444
8.10.2 Tracking in a Clutter Environment 449
8.10.3 Bearing-only Tracking with Passive Sonar 466
8.10.4 Stochastic Volatility Models 471
8.10.5 Fading Channels as Conditional Dynamic Linear Models 478
8.11 Exercises 486
References 487
Index 493
