|
Activity Number:
|
78
- Nonparametric Modeling
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Sunday, July 28, 2019 : 4:00 PM to 5:50 PM
|
|
Sponsor:
|
Section on Nonparametric Statistics
|
|
Abstract #304894
|
|
|
Title:
|
The Stationary Jackknife
|
|
Author(s):
|
Weilian Zhou* and Soumendra N Lahiri
|
|
Companies:
|
North Carolina State University and North Carolina State University
|
|
Keywords:
|
Resampling;
Jackknife;
Time series data;
Consistency
|
|
Abstract:
|
This work introduces a new resampling method in dealing with the variance estimation problem in dependent data situation. We regard the deleting length of the block as a random variable with the truncated geometric distribution, whereas the block length is fixed in the classical block jackknife method. We derive the expression of the bias and variance for the variance estimator for the sample mean of the dependent data. Then we prove the consistency of the estimator and generalize the proof to a larger range of estimators such as M-estimator. The optimal deleting block length is also illustrated in this work. The application of this method is in the time series data analysis and this method shows the ability that it is less sensitive to chosen length of the block when comparing with the block jackknife estimator.
|
Authors who are presenting talks have a * after their name.
