WILEY SERIES IN PROBABILITY AND STATISTICS
ESTABLISHED BY WALTER A. SHEWHART AND SAMUEL S. WILKS
Contents
Preface xv
1 Introduction 1
1.1 Classical and robust approaches to statistics 1
1.2 Mean and standard deviation 2
1.3 The “three-sigma edit” rule 5
1.4 Linear regression 7
1.4.1 Straight-line regression 7
1.4.2 Multiple linear regression 9
1.5 Correlation coefficients 11
1.6 Other parametric models 13
1.7 Problems 15
2 Location and Scale 17
2.1 The location model 17
2.2 M-estimates of location 22
2.2.1 Generalizing maximum likelihood 22
2.2.2 The distribution of M-estimates 25
2.2.3 An intuitive view of M-estimates 27
2.2.4 Redescending M-estimates 29
2.3 Trimmed means 31
2.4 Dispersion estimates 32
2.5 M-estimates of scale 34
2.6 M-estimates of location with unknown dispersion 36
2.6.1 Previous estimation of dispersion 37
2.6.2 Simultaneous M-estimates of location and dispersion 37
2.7 Numerical computation of M-estimates 39
2.7.1 Location with previously computed dispersion estimation 39
2.7.2 Scale estimates 40
2.7.3 Simultaneous estimation of location and dispersion 41
2.8 Robust confidence intervals and tests 41
2.8.1 Confidence intervals 41
2.8.2 Tests 43
2.9 Appendix: proofs and complements 44
2.9.1 Mixtures 44
2.9.2 Asymptotic normality of M-estimates 45
2.9.3 Slutsky’s lemma 46
2.9.4 Quantiles 46
2.9.5 Alternative algorithms for M-estimates 46
2.10 Problems 48
3 Measuring Robustness 51
3.1 The influence function 55
3.1.1 *The convergence of the SC to the IF 57
3.2 The breakdown point 58
3.2.1 Location M-estimates 58
3.2.2 Scale and dispersion estimates 59
3.2.3 Location with previously computed dispersion estimate 60
3.2.4 Simultaneous estimation 60
3.2.5 Finite-sample breakdown point 61
3.3 Maximum asymptotic bias 62
3.4 Balancing robustness and efficiency 64
3.5 *“Optimal” robustness 65
3.5.1 Bias and variance optimality of location estimates 66
3.5.2 Bias optimality of scale and dispersion estimates 66
3.5.3 The infinitesimal approach 67
3.5.4 The Hampel approach 68
3.5.5 Balancing bias and variance: the general problem 70
3.6 Multidimensional parameters 70
3.7 *Estimates as functionals 71
3.8 Appendix: proofs of results 75
3.8.1 IF of general M-estimates 75
3.8.2 Maximum BP of location estimates 76
3.8.3 BP of location M-estimates 76
3.8.4 Maximum bias of location M-estimates 78
3.8.5 The minimax bias property of the median 79
3.8.6 Minimizing the GES 80
3.8.7 Hampel optimality 82
3.9 Problems 84
4 Linear Regression 1 87
4.1 Introduction 87
4.2 Review of the LS method 91
4.3 Classical methods for outlier detection 94
4.4 Regression M-estimates 98
4.4.1 M-estimates with known scale 99
4.4.2 M-estimates with preliminary scale 100
4.4.3 Simultaneous estimation of regression and scale 103
4.5 Numerical computation of monotone M-estimates 103
4.5.1 The L1 estimate 103
4.5.2 M-estimates with smooth ψ-function 104
4.6 Breakdown point of monotone regression estimates 105
4.7 Robust tests for linear hypothesis 107
4.7.1 Review of the classical theory 107
4.7.2 Robust tests using M-estimates 108
4.8 *Regression quantiles 110
4.9 Appendix: proofs and complements 110
4.9.1 Why equivariance? 110
4.9.2 Consistency of estimated slopes under asymmetric errors 111
4.9.3 Maximum FBP of equivariant estimates 112
4.9.4 The FBP of monotone M-estimates 113
4.10 Problems 114
5 Linear Regression 2 115
5.1 Introduction 115
5.2 The linear model with random predictors 118
5.3 M-estimates with a bounded ρ-function 119
5.4 Properties of M-estimates with a bounded ρ-function 120
5.4.1 Breakdown point 122
5.4.2 Influence function 123
5.4.3 Asymptotic normality 123
5.5 MM-estimates 124
5.6 Estimates based on a robust residual scale 126
5.6.1 S-estimates 129
5.6.2 L-estimates of scale and the LTS estimate 131
5.6.3 Improving efficiency with one-step reweighting 132
5.6.4 A fully efficient one-step procedure 133
5.7 Numerical computation of estimates based on robust scales 134
5.7.1 Finding local minima 136
5.7.2 The subsampling algorithm 136
5.7.3 A strategy for fast iterative estimates 138
5.8 Robust confidence intervals and tests for M-estimates 139
5.8.1 Bootstrap robust confidence intervals and tests 141
5.9 Balancing robustness and efficiency 141
5.9.1 “Optimal” redescending M-estimates 144
5.10 The exact fit property 146
5.11 Generalized M-estimates 147
5.12 Selection of variables 150
……