Electronic Theses and Dissertations

Identifier

6719

Date

2021

Date of Award

7-13-2021

Document Type

Thesis

Degree Name

Master of Science

Major

Mathematical Sciences

Concentration

Statistics

Committee Chair

Su Chen

Committee Member

Lih-Yuan Deng

Committee Member

Wilfried J. J. Karmaus

Committee Member

Ching-Chi Yang

Abstract

Asthma is adisease causing wheezing, difficulty breathing, and coughing. The studies regarding whether childhood asthma is transmitted by maternal asthma through the alteration of DNA methylation are novel. Epigenetic studies work with a large pool of potential predictors while the sample size is much smaller, usually is a few of hundreds. Therefore, given the feature of ultra-high dimensionality, investigators demand to reduce the number of variables to perform designed statistical analyses. Feature-selection methods are often used as a screening step for detecting the small subset of variables associated with an outcome of interest. Many epigenetic studies used either the Training-Testing Screening (ttScreening) or the Recursive Random Forest (RRF), however, the comparison of those two methods has not been discussed. Therefore, in this thesis, we compared the performance of ttScreening and RRF with simulation studies and applied to a real data example in identifying potential CpG sites at birth of the newborns in the transmission of asthma from asthmatic mothers to childhood asthma.

Comments

Data is provided by the student.

Library Comment

dissertation or thesis originally submitted to the local University of Memphis Electronic Theses & dissertation (ETD) Repository.

Share

COinS