|
Activity Number:
|
371
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Tuesday, August 4, 2009 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Biometrics Section
|
| Abstract - #305635 |
|
Title:
|
Lack of Fit in Self-Modeling Regression
|
|
Author(s):
|
Lyndia C. Brumback*+
|
|
Companies:
|
University of Washington
|
|
Address:
|
Box 357232, Seattle, WA, 98195,
|
|
Keywords:
|
functional data ; mixed effects models ; arterial waveforms ; self-modeling
|
|
Abstract:
|
Self modeling regression is a method for modeling functional data. It assumes that each observed curve lies approximately on one typical curve, modeled with a nonparametric function, after separately transforming the x and y axes for the observed curve in a parametric manner. We show that when the typical curve is modeled with a natural regression spline and the curve-specific transformational parameters are modeled as random (Normal with mean zero), the model may suffer from lack of fit and the variance components may be estimated poorly. A random effects distribution that forces the realized curve-specific transformational parameters to have mean zero or the inclusion of a fixed transformational parameter improves estimation. We demonstrate the methods through simulations and application to pulse waveforms where one of the variance components represents blood pressure variability.
|
- 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 2009 program |
