New Advances in Statistics and Data Science (eBook)

Ding-Geng Chen, Zhezhen Jin, Gang Li, Yi Li, Aiyi Liu, Yichuan Zhao (Herausgeber)

eBook Download: PDF

2018 | 1st ed. 2017
XXIII, 348 Seiten
Springer International Publishing (Verlag)
978-3-319-69416-0 (ISBN)

Infos

This book is comprised of the presentations delivered at the 25^th ICSA Applied Statistics Symposium held at the Hyatt Regency Atlanta, on June 12-15, 2016. This symposium attracted more than 700 statisticians and data scientists working in academia, government, and industry from all over the world. The theme of this conference was the 'Challenge of Big Data and Applications of Statistics,' in recognition of the advent of big data era, and the symposium offered opportunities for learning, receiving inspirations from old research ideas and for developing new ones, and for promoting further research collaborations in the data sciences. The invited contributions addressed rich topics closely related to big data analysis in the data sciences, reflecting recent advances and major challenges in statistics, business statistics, and biostatistics. Subsequently, the six editors selected 19 high-quality presentations and invited the speakers to prepare full chapters for this book, which showcases new methods in statistics and data sciences, emerging theories, and case applications from statistics, data science and interdisciplinary fields. The topics covered in the book are timely and have great impact on data sciences, identifying important directions for future research, promoting advanced statistical methods in big data science, and facilitating future collaborations across disciplines and between theory and practice.

Ding-Geng Chen is a Fellow of the American Statistical Association and is currently the Wallace Kuralt distinguished professor at the University of North Carolina at Chapel Hill. He was a professor at the University of Rochester and the Karl E. Peace endowed eminent scholar chair in biostatistics at Georgia Southern University. He is also a senior statistics consultant for biopharmaceuticals and government agencies with extensive expertise in Monte-Carlo simulations, clinical trial biostatistics and public health statistics. Professor Chen has more than 100 referred professional publications, has co-authored and co-edited six books on clinical trial methodology, meta-analysis and public health applications, and has been invited nationally and internationally to give speeches on his research. Professor Chen was honored with the 'Award of Recognition' in 2014 by the Deming Conference Committee for highly successful advanced biostatistics workshop tutorials with his books.

 

Zhezhen Jin is a Professor of Biostatistics at Columbia University. His research interests in statistics include survival analysis, resampling methods, longitudinal data analysis, and nonparametric and semiparametric models. Dr. Jin has collaborated on research in the areas of cardiology, neurology, hematology, oncology and epidemiology. He is a founding editor-in-chief of Contemporary Clinical Trials Communications, serves as an associate editor for Lifetime Data Analysis, Contemporary Clinical Trials, and Communications for Statistical Applications and Methods, and is on the editorial board for Kidney International, the Journal of the International Society for Nephrology. Dr. Jin has published over 150 peer-reviewed research papers in statistical and medical journals, and is also a Fellow of the American Statistical Association.

 

Gang Li is a Professor of Biostatistics and Biomathematics at the University of California at Los Angeles (UCLA) and the Director of UCLA's Jonsson Comprehensive Cancer Center Biostatistics Shared Resource. His research interests include survival analysis, longitudinal data analysis, high dimensional and OMICs data analysis, clinical trials, and evaluation and development of biomarkers. He has published over 100 papers in a wide variety of prestigious journals such as the Annals of Statistics, the Journal of the American Statistical Association, the Journal of the Royal Statistical Society: Series B, Biometrika, and Biometrics. He is an Elected Fellow of the Institute of Mathematics, the American Statistical Association, and the Royal Statistical Society, as well as an Elected Member of the International Statistics Institute. He has been serving on the editorial board of several statistical journals including Biometrics. Dr. Li has been active in collaborating with researchers in basic science, translational research, and clinical trials, and has been a statistics Principal Investigator for multiple NIH funded projects.

 

Yi Li is a Professor of Biostatistics and Global Public Health at the University of Michigan School of Public Health (UM-SPH). He is currently the Director of China Initiatives at UM-SPH, and served as the Director of Kidney Epidemiology and Cost Center at the University of Michigan from 2011-2016. Dr. Li is an Elected Fellow of the American Statistical Association, and is serving as Associate Editor for several major statistical journals, including the Journal of the American Statistical Association, Biometrics, the Scandinavian Journal of Statistics, and Lifetime Data Analysis. His current research interests are survival analysis, longitudinal and correlated data analysis, measurement error problems, spatial models and clinical trial designs. He has published more than 140 papers in major statistical and biomedical journals, including the Journal of the American Statistical Association, the Journal of the Royal Statistical Society: Series B, Biometrika, Biometrics and the Proceedings of the National Academy of Sciences. His group has been developing methodologies for analyzing large-scale and high-dimensional datasets, with direct applications in observational studies as well in genetics/genomics. His methodologic research is funded by various NIH statistical grants starting from 2003. As principal investigator, Dr. Li has been leading a multi-year national project with focus on developing new measures to evaluate all dialysis facilities in the United States, with the goal of improving renal health care, saving lives and reducing cost. Dr. Li is actively involved in collaborative research in clinical trials and observational studies with researchers from the University of Michigan and Harvard University. The applications have included chronic kidney disease surveillance, organ transplantation, cancer preventive studies and cancer genomics.

  

Aiyi Liu is a Senior Investigator in the Biostatistics and Bioinformatics Branch of the Division of Intramural Population Health Research within the National Institutes of Health's Eunice Kennedy Shriver National Institute of Child Health and Human Development. A fellow of the American Statistical Association, Dr. Liu has authored/coauthored about 90 statistical methodological publications covering various topics including general statistical estimation theory, sequential methodology and adaptive designs, and statistical methods for diagnostic biomarkers. 

 

Yichuan Zhao is a Professor of Statistics at Georgia State University in Atlanta. His current research interest focuses on Survival Analysis, Empirical Likelihood Method, Nonparametric Statistics, Analysis of ROC Curves, Bioinformatics, Monte Carlo Methods, and Statistical Modeling of Fuzzy Systems. He has published over 70 research articles in Statistics and Biostatistics research fields. Dr. Zhao has organized the Workshop Series on Biostatistics and Bioinformatics since its initiation in 2012. He also organized the 25th ICSA Applied Statistics Symposium in Atlanta as a chair of the organizing committee to great success. He is currently serving as editor, or on the editorial board, for several statistical journals. Dr. Zhao was an elected member of the International Statistical Institute. 

Inhalt

Preface 6
Part I: Review and Theoretical Framework in Data Science (Chapters 1 –5) 7
Part II: Complex and Big Data Analysis (Chapters 6 –10) 7
Part III: Clinical Trials, Statistical Shape Analysis, and Applications (Chapters 11 –14) 8
Part IV: Statistical Modeling and Data Analysis (Chapters 15 –19) 9
Contents 12
Contributors 14
About the Editors 17
List of Chapter Reviewers 20
Part I Review and Theoretical Framework in Data Science 23
Statistical Distances and Their Role in Robustness 24
1 Introduction 24
2 The Discrete Setting 25
3 Chi-Squared Distance Measures 27
3.1 Loss Function Interpretation 28
3.2 Loss Analysis of Pearson and Neyman Chi-Squared Distances 31
3.3 Metric Properties of the Symmetric Chi-Squared Distance 32
3.4 Locally Quadratic Distances 38
4 The Continuous Setting 40
4.1 Desired Features 41
4.2 The L2- .4 -Distance 42
4.3 The Kolmogorov-Smirnov Distance 42
4.4 Exactly Quadratic Distances 44
5 Discussion 46
References 46
The Out-of-Source Error in Multi-Source Cross Validation-Type Procedures 48
1 Introduction 48
2 Literature Review and Notation 50
3 Framework 51
4 The OOS Error Estimation 52
4.1 Estimating the OOS Error 52
4.2 Bias and Variance of ?0 090d"0362 ??0
4.3 On Variance Estimation 56
5 Simulation Study 58
6 Discussion 60
Appendix 1: Some Useful Relations 61
Appendix 2: On Moments of Bivariate Normal Distribution 62
Appendix 3: Proofs 62
References 65
Meta-Analysis for Rare Events As Binary Outcomes 66
1 Introduction 66
2 Methods 68
2.1 Non-Parametric Meta-Analysis 68
2.1.1 Mantel-Haenszel Method 68
2.1.2 Peto Odds Ratio 69
2.1.3 Exact Method of Constructing Confidence Intervals for Risk Differences (Tian et al. 2009) 70
2.2 Parametric Meta-Analysis 71
2.2.1 Random-Effects Regression Model 71
2.2.2 Random-Effects Beta-Binomial Model 72
2.2.3 Random-Effects Poisson Model 72
2.3 Parametric Bootstrap Resampling Meta-Analysis 72
3 Case Studies 73
3.1 A Rosiglitazone Meta-Analysis Study 73
3.2 A Transplant Extrapolation Study with Everolimus 74
4 Summary 76
Appendix: Data for the rosiglitazone meta-analysis study from Nissen and Wolski (2007) 77
References 78
New Challenges and Strategies in Robust Optimal Design for Multicategory Logit Modelling 81
1 Introduction 81
2 Quantal Dose-Response Modelling 82
3 Confidence Regions and Intervals 84
4 Optimal Design Theory 84
5 Near-Optimal Robust Design Strategies 88
6 Discussion 92
References 93
Testing of Multivariate Spline Growth Model 95
1 Introduction 95
2 Modeling Growth with Smooth Functions 96
3 Testing of Mean Curves 98
3.1 Spline Approximation 98
3.2 Constructing a Test for Mean Spline Curves 99
4 Multivariate Spline Growth Curve Model 100
5 Computational Example: Modeling in Behavioral Cardiology 102
References 104
Part II Complex and Big Data Analysis 106
Uncertainty Quantification Using the Nearest Neighbor GaussianProcess 107
1 Introduction 107
2 Methods 109
2.1 Modeling with Gaussian Process 109
2.2 Bayesian Inference and Computational Considerations 112
3 Simulation Experiments 115
4 Application: Uncertainty Quantification for Surface Data 120
5 Conclusions and Discussion 122
References 123
Tuning Parameter Selection in the LASSO with UnspecifiedPropensity 126
1 Introduction 126
2 Model and Method 128
3 Tuning Parameter Selection in the LASSO 131
3.1 Multifold Cross Validation (CV) 131
3.2 Bayesian Information Criterion (BIC) 132
3.3 Variable Selection Stability (VSS) 133
3.4 Estimation Stability (ESCV) 133
4 Simulation Studies 134
5 Melanoma Study 137
6 Discussion 141
References 141
Adaptive Filtering Increases Power to Detect Differentially Expressed Genes 143
1 Introduction 143
2 Existing Filtering Methods 144
3 Proposed Method 145
4 A Data-Based Simulation Study 148
5 Conclusion 151
References 151
Estimating Parameters in Complex Systems with Functional Outputs: A Wavelet-Based Approximate Bayesian Computation Approach 153
1 Introduction 153
2 A Motivating Example: The Foliage-Echo Simulation System 157
3 Wavelet-Based Approximate Bayesian Computation 159
3.1 Review of Approximate Bayesian Computation 160
3.2 Wavelet Representation and Compression of Functional Data 161
3.3 A Gaussian Process Surrogate for the Simulator 163
3.4 Control the Uncertainty of Decision-Making in wABC Using GPS 166
4 The Algorithm and Parameter Settings 167
5 The Analysis of Simulated Foliage-Echo Data 168
6 Discussion 171
Appendix: More Details of the Foliage-Echo Simulator 172
References 174
A Maximum Likelihood Approach for Non-invasive Cancer Diagnosis Using Methylation Profiling of Cell-Free DNA from Blood 177
1 Introduction 177
2 Methods 179
2.1 Model the Methylation Probabilities 179
2.2 Estimate the Composition of Tumor-Derived cfDNA Using Methylation Data 181
2.3 Simulate the Methylation Sequencing Data of Plasma cfDNA Samples 183
2.4 Cancer Prediction Using Estimated Fraction of Tumor-Derived cfDNA and Evaluation Criteria 184
2.5 Applications to Real Data 185
3 Results 185
3.1 Estimation Accuracy Increases with Sequencing Depth and Fraction of Tumor-Derived cfDNA in Simulation Data 185
3.2 Estimated Fraction of Tumor-Derived cfDNA in Real Blood Samples Can Predict Normal from Liver Cancer Patients 187
3.3 The Fractions of Tumor-Derived cfDNA in the Blood of Liver Cancer Patients Are Significantly Decreased After Surgery 189
4 Discussion and Conclusions 190
References 191
Part III Clinical Trials, Statistical Shape Analysis and Applications 192
A Simple and Efficient Statistical Approach for Designing an Early Phase II Clinical Trial: Ordinal Linear Contrast Test 193
1 Introduction 193
2 Notation and Assumptions 194
2.1 Model Description 195
2.2 Monotonicity 195
2.3 Family-Wise Error Rate 196
3 Statistical Methods 196
3.1 Ordinal Linear Contrast Test 196
3.2 MCP-Mod 197
3.3 ANOVA F Test 198
3.4 MaxT Test 199
4 Dose Ranging 199
5 Method Comparisons 201
5.1 OLCT Approach 201
5.2 MCP-Mod Approach 202
5.3 ANOVA Approach 203
5.4 MaxT Approach 203
5.5 Comparisons 204
5.6 When to Use MCP-Mod and When Not? 206
5.7 When to Use OLCT and When Not? 207
5.8 Limitations of ANOVA F Test 208
6 Discussion 209
References 210
Landmark-Constrained Statistical Shape Analysis of Elastic Curves and Surfaces 211
1 Introduction 211
2 Landmark-Constrained Shape Analysis 214
2.1 Unconstrained Representation Spaces of Curves and Surfaces 215
2.2 Landmark-Constrained Shape Space for Curves 216
2.3 Landmark-Constrained Shape Space for Surfaces 217
2.4 Motivating Examples 219
2.5 Additional Examples 221
3 Statistical Analysis of Landmark-Constrained Shapes 223
3.1 Sample Averaging 223
3.1.1 Examples 224
3.2 Summarization of Variability 225
3.2.1 Examples 226
4 Summary 227
References 228
Phylogeny-Based Kernels with Application to Microbiome Association Studies 231
1 Introduction 231
2 Methods 234
2.1 Phylogeny-Induced Correlation Structure Among OTUs 234
2.2 A Phylogeny-Based Kernel for Microbiome Data 236
2.3 Kernel-Machine (KM) Association Test 237
2.3.1 Single Kernel-Based KM Association Test 238
2.3.2 Multiple Kernel-Based Optimal KM Association Test 240
3 Simulation Studies 240
3.1 Simulation Details 241
3.2 Results on Simulated Data 242
4 Application to a Real Data Set 242
5 Discussion 246
References 248
Accounting for Differential Error in Time-to-Event Analyses Using Imperfect Electronic Health Record-Derived Endpoints 252
1 Introduction 252
2 Methods 255
2.1 Definitions and Notation 256
2.2 Discrete Proportional Hazards Model 256
2.3 Adjustment for Error in Event Times 257
2.4 Incorporating Person-Level Validation Data 258
2.5 Differential Error in Event Times 260
2.6 Simulation Study Design 260
3 Results 262
3.1 Non-differential Error in Dates 262
3.2 Differential Error in Dates 263
4 Discussion 265
References 267
Part IV Statistical Modeling and Data Analysis 269
Modeling Inter-Trade Durations in the Limit Order Market 270
1 Introduction 270
2 Empirical Facts 273
3 Model and Estimation 275
3.1 Model Specification 275
3.2 Maximum Likelihood 277
4 Empirical Analysis 279
5 Conclusion 286
References 286
Assessment of Drug Interactions with Repeated Measurements 288
1 Introduction 288
2 Median Effect Model for Drug Combination Effects 289
2.1 Median Effect Model 290
2.2 Ray Design 290
2.3 Loewe Additivity 291
2.4 Drug Combination Effects with Repeated Measurements 292
3 Confidence Interval Estimation at the Observed Combination 293
4 Confidence Bound for Interaction Index on a Fixed Ray 295
5 Simulation Study 296
6 Application 299
7 Discussion 299
References 301
Statistical Indices for Risk Tracking in Longitudinal Studies 303
1 Introduction 303
2 Rank-Based Tracking Indices 305
2.1 Rank-Tracking Probabilities 306
2.2 Rank-Tracking Probability Ratios 307
2.3 Mean-Integrated RTPs and RTPRs 308
3 Estimation and Inference Methods 309
3.1 Nonparametric Mixed Models and Prediction 309
3.2 Estimation of Tracking Indices 311
3.3 Bootstrap Confidence Intervals 313
4 Application to the NGHS Data 313
4.1 Rank-Tracking for BMI 313
4.2 Rank-Tracking for SBP 315
5 Simulation 316
6 Discussion 318
References 319
Statistical Analysis of Labor Market Integration: A Mixture Regression Approach 322
1 Introduction 322
2 Methods 323
2.1 Data 323
2.2 Multivariate Binary Mixture 324
3 Analysis 325
3.1 Normal Life-Course 326
3.2 Weak Labor Market Integration 328
4 Concluding Remarks 328
Appendix 329
References 330
Bias Correction in Age-Period-Cohort Models Using Eigen Analysis 331
1 Introduction 331
2 Age-Period-Cohort Model and the Identification Problem 333
2.1 The Age-Period-Cohort Model 333
2.2 The Identification Problem 334
2.3 The Intrinsic Estimator Addressing the Identification Problem 336
3 The Bias Correction Method 337
3.1 Bias Correction Algorithm 338
3.2 Standard Error Estimation 338
3.3 Algorithm for Standard Error Estimation After Bias Correction 339
4 Application 340
4.1 Cervical Cancer Incidence Rate 340
4.2 Chronic Obstructive Pulmonary Disease Mortality 343
4.3 Importance of the Variance/Standard Error Estimation 345
5 Conclusion and Discussion 346
References 348
Index 350

Produkt-Details

• Erscheint lt. Verlag: 17.1.2018
• Reihe/Serie: ICSA Book Series in Statistics
• Reihe/Serie: ICSA Book Series in Statistics
• Zusatzinfo: XXIII, 348 p. 74 illus., 41 illus. in color.
• Verlagsort: Cham
• Sprache: englisch
• Themenwelt: Mathematik / Informatik ► Mathematik ► Wahrscheinlichkeit / Kombinatorik
• Themenwelt: Medizin / Pharmazie ► Allgemeines / Lexika
• Themenwelt: Technik
• Themenwelt: Wirtschaft
• Schlagworte: Big Data • Clinical trials design • DNA statistical analysis • functional data analysis • Gene expression analysis • high dimensional statistical method • Longitudinal data analysis • Nonparametric Statistics • phylogeny-based kernels • spline growth model • statistical genetics and bioinformatics • Statistical Methods • statistical shape analysis • survival data analysis • uncertainty quantification
• ISBN-10: 3-319-69416-2 / 3319694162
• ISBN-13: 978-3-319-69416-0 / 9783319694160