Online Program Home
  My Program

All Times EDT

Abstract Details

Activity Number: 339 - Individual Treatment Rule and Precision Medicine
Type: Contributed
Date/Time: Wednesday, August 5, 2020 : 10:00 AM to 2:00 PM
Sponsor: Biometrics Section
Abstract #313930
Title: Statistical Method for Small-Sample Inferences of Multi-Way Data
Author(s): Bin Guo* and Baolin Wu and Lynn Eberly
Companies: University of Minnesota and University of Minnesota and University of Minnesota
Keywords: Linear mixed model; small sample ; multi-way

High-throughput technologies are being increasingly used to generate a multitude of related phenotypes for patients to study their clinical impacts. For example, in a study of neurodegenerative disease, we can collect data for a sample of subjects across a series of time points and brain regions. Most existing methods for analyzing these multi-way data mainly focus on exploring common patterns across multiple dimensions, such as the widely used factor decomposition methods. They are not designed to directly test and investigate the covariate effects. Also, existing methods based on asymptotic approximation often do not perform well for small-sample studies, which are common for neuroimaging studies. To tackle these issues, we proposed a general joint statistical framework based on a multivariate mixed effect model for analyzing multi-way data that can readily incorporate covariate information and draw accurate and efficient inference for small sample studies. We conduct simulation studies to show type I errors are well-controlled. We demonstrate the utility of proposed methods with application to a neuroimaging dataset.

Authors who are presenting talks have a * after their name.

Back to the full JSM 2020 program