|
Activity Number:
|
468
- Modern Topics in Hypothesis Testing
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Thursday, August 6, 2020 : 10:00 AM to 2:00 PM
|
|
Sponsor:
|
IMS
|
|
Abstract #313207
|
|
|
Title:
|
The Robustness of Covtest in Exponential Family
|
|
Author(s):
|
Dewei Zhong* and John Kolassa
|
|
Companies:
|
and Rutgers University
|
|
Keywords:
|
|
|
Abstract:
|
Lockhart [1] suggest the covtest as a method to test the significance of a single variable in regression. Asymptotic results of exponential distribution were proposed to characterize the distribution of the statistic, but the asymptotic results are known to be inaccurate. We propose the gamma distribution is a better fit for covtest statistic than the exponential one. The order of error between the covtest statistic's moments and the asymptotic approximation is found. And we find the covtest can be used even if the error term of regression does not follow normal distribution. We show it is valid for the Laplace error term as well.
|
Authors who are presenting talks have a * after their name.
Back to the full JSM 2020 program
|
