Perturbation and scaled Cook's distance

DOI10.1214/12-AOS978zbMATH Open1273.62180arXiv1204.0064OpenAlexW1990560976WikidataQ41234164 ScholiaQ41234164MaRDI QIDQ447828FDOQ447828

Hyunsoon Cho, Joseph G. Ibrahim, Hongtu Zhu

Publication date: 29 August 2012

Published in: The Annals of Statistics (Search for Journal in Brave)

Abstract: Cook's distance [Technometrics 19 (1977) 15-18] is one of the most important diagnostic tools for detecting influential individual or subsets of observations in linear regression for cross-sectional data. However, for many complex data structures (e.g., longitudinal data), no rigorous approach has been developed to address a fundamental issue: deleting subsets with different numbers of observations introduces different degrees of perturbation to the current model fitted to the data, and the magnitude of Cook's distance is associated with the degree of the perturbation. The aim of this paper is to address this issue in general parametric models with complex data structures. We propose a new quantity for measuring the degree of the perturbation introduced by deleting a subset. We use stochastic ordering to quantify the stochastic relationship between the degree of the perturbation and the magnitude of Cook's distance. We develop several scaled Cook's distances to resolve the comparison of Cook's distance for different subset deletions. Theoretical and numerical examples are examined to highlight the broad spectrum of applications of these scaled Cook's distances in a formal influence analysis.

Full work available at URL: https://arxiv.org/abs/1204.0064

Recommendations

• Structured Perturbations Part II: Componentwise Distances
• Structured Perturbations Part I: Normwise Distances
• Exact distribution of Cook's distance and identification of influential observations
• A NOTE ON LARGE-SAMPLE PROPERTY OF COOK'S DISTANCE
• Cook's distance in local polynomial regression
• Partial distance correlation with methods for dissimilarities
• Distance matrices perturbed by Laplacians.
• The frechet distance revisited and extended

zbMATH Keywords

perturbationCook's distanceconditionally scaled Cook's distancerelative influentialscaled Cook's distancesize issue

Mathematics Subject Classification ID

Diagnostics, and linear inference and regression (62J20) Generalized linear models (logistic models) (62J12)

Cites Work

• Title not available (Why is that?)
• Maximum Likelihood Estimation of Misspecified Models
• Robust Statistics
• Regression Model Diagnostics
• Title not available (Why is that?)
• Title not available (Why is that?)
• Deletion diagnostics for generalised estimating equations
• Estimation When a Parameter is on a Boundary
• Case-deletion measures for models with incomplete data
• Title not available (Why is that?)
• Title not available (Why is that?)
• Detection of Influential Observation in Linear Regression
• Title not available (Why is that?)
• Perturbation selection and influence measures in local influence analysis
• A Simple Derivation of Deletion Diagnostic Results for the General Linear Model with Correlated Errors
• Outlier..........s
• Analysis of Infant Growth Curves Using Multivariate Adaptive Splines
• Influence Diagnostics and Outlier Tests for Semiparametric Mixed Models
• Exponential family nonlinear models
• Influence Diagnostics for Linear Longitudinal Models
• Case-Deletion Diagnostics for Mixed Models
• On Wielandt's inequality and its application to the asymptotic distribution of the eigenvalues of a random symmetric matrix
• Asymptotics for estimation and testing procedures under loss of identifiability
• Influence analysis based on the case sensitivity function
• Model-Checking Techniques Based on Cumulative Residuals
• Cook's distance in linear longitudinal models

Cited In (8)

• Case deletion diagnostics for GMM estimation
• A NOTE ON LARGE-SAMPLE PROPERTY OF COOK'S DISTANCE
• High-dimensional influence measure
• Diagnostic measures for functional linear model with nonignorable missing responses
• A robust variable screening method for high-dimensional data
• Multiple Influential Point Detection in High Dimensional Regression Spaces
• Local influence analysis for GMM estimation
• Estimating large-scale general linear and seemingly unrelated regressions models after deleting observations

This page was built for publication: Perturbation and scaled Cook's distance

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q447828)