Activity Number:
|
212
|
Type:
|
Contributed
|
Date/Time:
|
Tuesday, August 13, 2002 : 10:30 AM to 12:20 PM
|
Sponsor:
|
Biometrics Section*
|
Abstract - #300669 |
Title:
|
Detecting Multicollinearity in Logistic Regression Models
|
Author(s):
|
Ariful Haque*+ and Abbas Jawad and Mayadah Shabbout
|
Affiliation(s):
|
Children's Hospital, Philadelphia and Children's Hospital, Philadelphia and Children's Hospital, Philadelphia
|
Address:
|
3535 Market Street, Suite 1032, Philadelphia, Pennsylvania, 19104, USA
|
Keywords:
|
Multicollinearity ; Singular Value Decomposition ; Condition Index ; Variance Decomposition Proportions
|
Abstract:
|
This paper deals with the examination and detection of multicollinearity in logistic regression models. Belsley, Kuh, and Welsch (1980) suggested the diagnostic measures of condition indices and variance decomposition proportions for detecting multicollinearity in multiple linear regression. Here, we generalize these measures for detecting possible dependencies among the covariates in logistic regression models. In this paper, we show that our generalization is capable of detecting collinearity for logistic models by assessing the extent to which each of the logistic regression parameters is degraded by the presence of such collinear relations. This approach can be used for other type of generalized linear models since logistic regression is a class of generalized linear models. We illustrate the methodology with pediatric clinical data.
|
- The address information is for the authors that have a + after their name.
- Authors who are presenting talks have a * after their name.
Back to the full JSM 2002 program |