|
Activity Number:
|
510
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Thursday, August 2, 2007 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Section on Statistical Computing
|
| Abstract - #310162 |
|
Title:
|
Maximum Likelihood--Based Confidence Intervals for Complementary Poisson Rate Parameters with Data Subject to Misclassification
|
|
Author(s):
|
Kent Riggs*+
|
|
Companies:
|
Stephen F. Austin State University
|
|
Address:
|
Box 13040 SFA Station, Nacogdoches, TX, 75962-3040,
|
|
Keywords:
|
asymptotic intervals ; misclassification ; double sampling ; simulation
|
|
Abstract:
|
Of interest is statistical inference for two complementary Poisson rate parameters where the data is subject to misclassification. A double sampling scheme is employed to account for misclassification and to estimate the binomial misclassification parameters. We propose an EM algorithm to find restricted and unrestricted MLEs. Next, we derive the maximum-likelihood-based confidence intervals for the Poisson and binomial parameters using the Wald, score, and likelihood ratio statistics. We then study coverage and width properties of the three asymptotic intervals. Finally, we apply the confidence intervals to a real-life data set.
|
