|
Activity Number:
|
508
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Thursday, August 10, 2006 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Section on Nonparametric Statistics
|
| Abstract - #307053 |
|
Title:
|
Robust Nonparametric Confidence Intervals and Tests for the Median in the Presence of (c,r)-Contamination
|
|
Author(s):
|
Masakazu Ando*+ and Itsuro Kakiuchi and Miyoshi Kimura
|
|
Companies:
|
Japan Society for the Promotion of Science and Kobe University and Nanzan University
|
|
Address:
|
Nagoya City University, Nagoya, 467-8501, Japan
|
|
Keywords:
|
robust nonparametric inference ; median ; confidence interval ; sign test ; coverage probability ; (c,r)-contamination
|
|
Abstract:
|
The problem of constructing robust nonparametric confidence intervals and tests for the median is considered when the data distribution is unknown and the data may be contaminated. A new form of the (c,r)-neighborhood is proposed and it is used in order to describe the contamination of the data. The (c,r)-neighborhood is a generalization of the neighborhoods defined in terms of epsilon-contamination and total variation distance. A modification of the sign test and its associated confidence intervals are proposed, and their robustness and efficiency are studied under the (c,r)-neighborhood of a continuous distribution. The derived results are natural extensions of those in the case of epsilon-contamination. Some tables and figures of coverage probability and expected length for the confidence intervals are given. Further possible generalizations are also discussed.
|
