Bayesian Variable Selection in High-dimensional Applications

Abstract

Advances in research technologies over the past few decades have encouraged the proliferation of massive datasets, revolutionizing statistical perspectives on high-dimensionality. Highthroughput technologies have become pervasive in diverse scientific disciplines and continued to generate data of increasingly complex phenomena, altering the course of statistical developments both in methodology and theory. A major focus of the intensive methodological research has centered around variable selection, which has become fundamental to knowledge extraction from such challenging data. The problem of variable selection refers to the statistical endeavor of selecting a subset of observed characteristics, which collectively provide a good description of an observed phenomenon. Of particular interest are settings where such a subset is parsimonious.