|
Activity Number:
|
407
|
|
Type:
|
Topic Contributed
|
|
Date/Time:
|
Wednesday, August 1, 2007 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Business and Economics Statistics Section
|
| Abstract - #310112 |
|
Title:
|
A Local Goodness-of-Fit Diagnostic Based on the Log Determinant of the Sample Covariance Matrix
|
|
Author(s):
|
Tucker S. McElroy and Scott Holan*+
|
|
Companies:
|
U.S. Census Bureau and University of Missouri-Columbia
|
|
Address:
|
3601 W Broadway Apt 30303, Columbia, MO, 65203,
|
|
Keywords:
|
Frequency Domain ; Nonstationary ; Time Series
|
|
Abstract:
|
The log determinant of the covariance matrix of a stationary process has been used by various authors to provide information, and in particular can be used as a goodness-of-fit diagnostic on the residual sample autocovariances. This is equivalent to considering a total integral of the logged residual spectral density overall frequencies. We consider generalized goodness-of-fit measures that integrate over narrow bands of frequencies for use in assessing local goodness of fit. Additionally, we present theoretical and empirical results demonstrating the useful properties of the diagnostic. Finally, we illustrate the effectiveness of our diagnostic through application to real data from the U.S. Census Bureau.
|
