< Back to previous page

Project

Essays on the elimination of high-dimensional nuisance parameters

The aim of this research is to develop methods for the elimination of nuisance parameters in high-dimensional settings. The high dimensionality is a key ingredient in data-rich applications and can come from a variety of sources. For example, trade data are indexed by origin, destination, time, industrial sector, and possibly subdivisions thereof. Each of these indices may carry its own set of nuisance parameters to control for effects of origin, destination, and so on, resulting in a very large numbers of nuisance parameters. A more abstract example is the situation where every unit of observation carries its own nuisance parameter (in some model), and one is interested in certain functionals of the ensemble of nuisance parameters. The elimination of high-dimensional nuisance parameters (or the estimation of functionals thereof) is statistically challenging in most models because the classic methods typically all fail. Solutions to the problem, exact or approximate, will be sought from several angles: the neo-Fisherian modified-likelihood perspective (with several variants); the Bayesian integrated-likelihood approach; and empirical-likelihood methods (for semi-parametric problems).

Date:22 Jun 2018 →  22 Jun 2022
Keywords:nuisance parameters
Disciplines:Applied economics, Economic history, Macroeconomics and monetary economics, Microeconomics, Tourism
Project type:PhD project