|
Activity Number:
|
543
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Thursday, August 2, 2007 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Section on Nonparametric Statistics
|
| Abstract - #308587 |
|
Title:
|
Functional Density Warping
|
|
Author(s):
|
Zhen Zhang*+ and Hans-Georg Müller
|
|
Companies:
|
University of California, Davis and University of California, Davis
|
|
Address:
|
112 Koshland Way, Santa Cruz, CA, 95064,
|
|
Keywords:
|
Age-at-death density ; Hazard rate ; Functional data analysis ; Time warping ; Quantile function ; Kernel smoothing
|
|
Abstract:
|
When estimating an overall density function based on observations on individual densities for a sample of cohorts, varying scales between cohorts can be a problem which is ignored by conventional density estimation. The proposed density warping approach incorporates varying time scales between cohorts, yielding an overall density that more adequately represents the sample of cohort densities. Observed cohort densities are assumed to be time-warped realizations of an underlying stochastic process. In a first step, we obtain time-synchronized distribution functions from a warping mapping, from which both overall density and hazard functions are derived. Asymptotic properties of the resulting density estimates are derived, and we illustrate their performance in simulations and for longevity data obtained for 142 cohorts of Mexican fruit flies.
|
