Skip to Main Content
Frequently Asked Questions
Submit an ETD
Global Search Box
Need Help?
Keyword Search
Participating Institutions
Advanced Search
School Logo
Files
File List
Thesis.pdf (7.78 MB)
ETD Abstract Container
Abstract Header
Nonparametric Covariance Estimation for Longitudinal Data
Author Info
Blake, Tayler Ann, Blake
ORCID® Identifier
http://orcid.org/0000-0003-2275-9600
Permalink:
http://rave.ohiolink.edu/etdc/view?acc_num=osu15256491898913
Abstract Details
Year and Degree
2018, Doctor of Philosophy, Ohio State University, Statistics.
Abstract
Estimation of an unstructured covariance matrix is difficult because of the challenges posed by parameter space dimensionality and the positive definiteness constraint that estimates should satisfy. We propose a general framework for nonparametric covariance estimation for longitudinal data where the variables have a natural ordering. Modeling the Cholesky decomposition of the covariance matrix removes constraints from estimation, including those posed by positive definiteness. In addition, the Cholesky decomposition enjoys the added advantage over alternative matrix decompositions by supplying a meaningful statistical interpretation of the corresponding estimated parameters. We illustrate the equivalence of covariance estimation and the estimation of a varying coefficient autoregressive model. By defining the varying coefficient as a bivariate function, we naturally accommodate sparsely or irregularly sampled longitudinal data without the need for imputation. This framework extends the set of tools available for covariance estimation to any of those employed in the typical function estimation setting. Viewing stationarity as a form of simplicity or parsimony in covariance models, we specify the varying coefficient as a function so that we can conveniently penalize the components capturing the nonstationarity in the fitted function. Casting covariance estimation as bivariate smoothing problem, we demonstrate construction of a covariance estimator using the smoothing spline framework and a penalized B-spline expansion. A simulation study establishes the advantage of our estimator over alternative estimators proposed in this setting. We analyze a longitudinal dataset to illustrate application of the methodology and compare our estimates to those resulting from alternative models proposed for the covariance for longitudinal data.
Committee
Yoonkyung Lee, PhD (Advisor)
Pages
163 p.
Subject Headings
Statistics
Keywords
longitudinal data, Hilbert space, nonparametric function estimation, covariance estimation
Recommended Citations
Refworks
EndNote
RIS
Mendeley
Citations
Blake, Blake, T. A. (2018).
Nonparametric Covariance Estimation for Longitudinal Data
[Doctoral dissertation, Ohio State University]. OhioLINK Electronic Theses and Dissertations Center. http://rave.ohiolink.edu/etdc/view?acc_num=osu15256491898913
APA Style (7th edition)
Blake, Blake, Tayler.
Nonparametric Covariance Estimation for Longitudinal Data.
2018. Ohio State University, Doctoral dissertation.
OhioLINK Electronic Theses and Dissertations Center
, http://rave.ohiolink.edu/etdc/view?acc_num=osu15256491898913.
MLA Style (8th edition)
Blake, Blake, Tayler. "Nonparametric Covariance Estimation for Longitudinal Data." Doctoral dissertation, Ohio State University, 2018. http://rave.ohiolink.edu/etdc/view?acc_num=osu15256491898913
Chicago Manual of Style (17th edition)
Abstract Footer
Document number:
osu15256491898913
Download Count:
608
Copyright Info
© 2018, some rights reserved.
Nonparametric Covariance Estimation for Longitudinal Data by Tayler Ann Blake Blake is licensed under a Creative Commons Attribution-NonCommercial-NoDerivs 3.0 Unported License. Based on a work at etd.ohiolink.edu.
This open access ETD is published by The Ohio State University and OhioLINK.