|
Activity Number:
|
386
- Nonparametric Modeling II
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Thursday, August 12, 2021 : 12:00 PM to 1:50 PM
|
|
Sponsor:
|
Section on Nonparametric Statistics
|
|
Abstract #318814
|
|
|
Title:
|
Data Harmonization via Regularized Nonparametric Mixing Distribution Estimation
|
|
Author(s):
|
Steven Wilkins-Reeves* and Yen-Chi Chen and Kwun Chuen Gary Chan
|
|
Companies:
|
University of Washington, Department of Statistics and University of Washington and University of Washington
|
|
Keywords:
|
Nonparametric;
Mixing Distribution;
EM Algorithm;
Data Harmonization;
Alzheimer’s ;
Dementia
|
|
Abstract:
|
Data harmonization is the process by which an equivalence is developed between two variables measuring a common trait. Our problem is motivated by dementia research that many tests may measure the same underlying cognitive ability such as language or memory. We connect this rarely studied problem in statistics to mixing distribution estimation. One of the fundamental challenges of mixing distribution estimation is the non-identifiability problem. We develop a simple regularized method which enforces uniqueness of the maximum likelihood estimator, and show how a functional EM algorithm will converge weakly to the regularized maximum likelihood estimator. Additionally, we develop goodness of fit tests for the mixing likelihood which is an area neglected in most mixing distribution estimation problems. We apply our method to the National Alzheimer’s Coordination Center Uniform Dataset and show that we can use our method to convert between score measurements and capture the measurement error. We show that this method outperforms standard techniques used in modern dementia research.
|
Authors who are presenting talks have a * after their name.
