AMUSE

Date: October 14, 2019

Time: 6:00PM - 7:00PM

Location: BLOC 220

Speaker: Adrian Thompson, Dept of Physics, Texas A&M University

Title: Copulas and their Applications to Bayesian Analysis in Physics

Abstract: Many data-driven fields such as finance, meteorology, engineering and physics often encounter data with a high number of dimensions. Modeling multivariate data, even with low dimensionality, can be challenging. The copula is a statistical object that separately combines the correlations and the one-dimensional projections of a dataset into one entity. This property provides an effective way of modeling multivariate data that scales well with dimensionality. Another topic, Bayesian analysis, is used frequently in physics to estimate the likelihood of a particular measurement given data. I will discuss the copula and its applications to Bayesian analysis in neutrino physics, an explosively growing field over the past decade, which comes with a large number of physical parameters.