Defense Date

1994

Document Type

Dissertation

Degree Name

Doctor of Philosophy

Department

Biostatistics

First Advisor

Daijin Ko

Second Advisor

Pippa M. Simpson

Abstract

The standardized influence matrix is a generalization of the standardized influence function and Cook’s approach to local influence. It provides a general and unified approach to judge the suitability of statistical inference based on parametric models. It characterizes the local influence of data deviations from parametric models on various estimators, including generalized linear models. Its use for both robustness measures and diagnostic procedures has been studied. With global robust estimators, diagnostic statistics are proposed and shown to be useful in detecting influential points for linear regression and logistic regression models. Robustness of various estimators is compared via. the standardized influence matrix and a new robust estimator for logistic regression models is presented.

Comments

Scanned, with permission from the author, from the original print version, which resides in University Archives.

Rights

© The Author

Is Part Of

VCU University Archives

Is Part Of

VCU Theses and Dissertations

Date of Submission

6-29-2017

Included in

Biostatistics Commons

Share

COinS