Empirical Processes in M-Estimation (Cambridge Series in Statistical and Probabilistic Mathematics, Series Number 6) [Illustrated]
0521123259, 9780521123259
The theory of empirical processes provides valuable tools for the development of asymptotic theory in (nonparametric) st
Table of contents : Cover Applications of Empirical Process Theory Series Title Copyright Contents Preface Guide to the Reader 1 Introduction 1.1. Some examples from statistics 1.2. Problems and complements 2 Notation and Definitions 2.1. Stochastic order symbols 2.2. The empirical process 2.3. Entropy 2.4. Examples 2.5. Notes 2.6. Problems and complements 3 Uniform Laws of Large Numbers 3.1. Uniform laws of large numbers under finite entropy with bracketing 3.2. The chaining technique 3.3. A maximal inequality for weighted sums 3.4. Symmetrization 3.5. Hoeffding's inequality 3.6. Uniform laws of large numbers under random entropy conditions 3.7. Examples 3.8. Notes 3.9. Problems and complements 4 First Applications: Consistency 4.1. Consistency of maximum likelihood estimators 4.2. Examples 4.3. Consistency of least squares estimators 4.4. Examples 4.5. Notes 4.6. Problems and complements 5 Increments of Empirical Processes 5.1. Random entropy numbers and asymptotic equicontinuity 5.2. Random entropy numbers and classes depending on n 5.3. Empirical entropy and empirical norms 5.4. A uniform inequality based on entropy with bracketing 5.5. Entropy with bracketing and asymptotic equicontinuity 5.6. Modulus of continuity 5.7. Entropy with bracketing and empirical norms 5.8. Notes 5.9. Problems and complements 6 Central Limit Theorems 6.1. Definitions 6.2. Sufficient conditions for 3 to be P-Donsker 6.3. Useful theorems 6.4. Measurability 6.5. Notes 6.6. Problems and complements 7 Rates of Convergence for Maximum Likelihood Estimators 7.1. The main idea 7.2. An exponential inequality for the maximum likelihood estimator 7.3. Convex classes of densities 7.4. Examples 7.5. Notes 7.6. Problems and complements 8 The Non-I.I.D. Case 8.1. Independent non-identically distributed random variables 8.1.1. Maximal inequalities for weighted sums revisited 8.2. Martingales 8.3. Application to maximum likelihood 8.4. Examples 8.5. Notes 8.6. Problems and complements 9 Rates of Convergence for Least Squares Estimators 9.1. Sub-Gaussian errors 9.2. Errors with exponential tails 9.3. Examples 9.4. Notes 9.5. Problems and complements 10 Penalties and Sieves 10.1. Penalized least squares 10.2. Penalized maximum likelihood 10.2.1. Roughness penalty on the density 10.2.2. Roughness penalty on the log-density 10.3. Least squares on sieves 10.4. Maximum likelihood on sieves 10.5. Notes 10.6. Problems and complements 11 Some Applications to Semiparametric Models 11.1. Partial linear models 11.2. Mixture models 11.2.1. Introduction 11.3. A single-indexed model with binary explanatory variable 11.4. Notes 11.5. Problems and complements 12 M-Estimators 12.1. Introduction 12.2. Estimating a regression function using a general loss function 12.3. Classes of functions indexed by a finite-dimensional parameter 12.3.1. Least squares 12.3.2. Maximum likelihood 12.3.3. Asymptotic normality 12.4. Notes 12.5. Problems and complements Appendix References Symbol Index Author Index Subject Index