Principal wave analysis for high-dimensional structured data with applications to epigenomics and neuroimaging studies

High-dimensional structured data are emerging and accumulating in biomedical research fields. Examples include epigenomics and neuroimaging studies. In these studies, it is often required to extract biologically meaningful patterns and identify relevant biological features from highdimensional structured data. Motivated by this problem, we propose a new statistical learning method named Principal Wave Analysis (PWA). The practical merits of PWA are shown through simulation studies incorporating diverse types of signal patterns as well as its applications to epigenomic and neuroimaging data.

Keywords

dimension reduction, feature selection, multiresolution analysis, wavelets

The full text of this article is unavailable through your IP address: 52.14.26.141

Received 30 October 2019

Accepted 11 December 2020

Published 11 January 2022