|
Activity Number:
|
75
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Sunday, August 6, 2006 : 4:00 PM to 5:50 PM
|
|
Sponsor:
|
Section on Nonparametric Statistics
|
| Abstract - #306578 |
|
Title:
|
A Graphical Method for Testing the Equality of Regression Curves
|
|
Author(s):
|
Kee-Hoon Kang*+ and Cheolwoo Park
|
|
Companies:
|
Hankuk University of Foreign Studies and University of Georgia
|
|
Address:
|
Wangsan, Yongin, 449-791, South Korea
|
|
Keywords:
|
sizer ; nonparametric regression ; comparison of regression curves
|
|
Abstract:
|
One of the most important problems of statistical inference is the comparison of two or more groups, which occurs in a variety of contexts. Comparison of population curves in the nonparametric way includes densities, regression curves, survival functions and some other characteristic functions of the variable of interest. In this article we introduce a new graphical method for testing the equality of several independent regression curves. Our method is based on analyzing SiZer (SIgnificant ZERo crossing of the differences) maps, which are scale-space based visualization tools for statistical inference. We use SiZer for comparing residual distributions under the null hypothesis. The proposed method is tested under various simulation settings and real examples.
|
