Optimal Conditionally Unbiased Bounded-Influence Inference in Dynamic Location and Scale Models

• Published in: Journal of the American Statistical Association Vol. 100; no. 470; pp. 628 - 641
• Main Authors: Mancini, Loriano, Ronchetti, Elvezio, Trojani, Fabio
• Format: Journal Article
• Language: English
• Published: Alexandria, VA Taylor & Francis 01.06.2005; American Statistical Association; Taylor & Francis Ltd
• Subjects: Approximation; Estimators; Exact sciences and technology; Inference; Inference from stochastic processes; time series analysis; Influence function; M-estimator; Mathematics; Maximum likelihood estimation; Modeling; Monte Carlo simulation; Nonparametric inference; Normal distribution; Parametric inference; Parametric models; Probability and statistics; Robust estimation and testing; Scale modeling; Sciences and techniques of general use; Simulations; Statistical inference; Statistics; Theory and Methods; Time series; Time series model; Time series models; Unbiased estimators; Robust estimation and resting; Conditional distribution; Estimator robustness; M-estimator; Robust estimation; Optimal estimation; Numerical approximation; Scale parameter; Consistent estimator; Pseudolikelihood; Dynamic model; Stationary series; Laplace approximation; Monte Carlo method; Test robustness; Time series; Location parameter; Statistical estimation; Time series model; Heteroscedasticity; Statistical method; Influence function; Unbiased estimation; Non linear model; Maximum likelihood; M estimation; Parametric test; Autoregressive processes

This article studies the local robustness of estimators and tests for the conditional location and scale parameters in a strictly stationary time series model. We first derive optimal bounded-influence estimators for such settings under a conditionally Gaussian reference model. Based on these result...