Utility Based Model Selection and Model Averaging

28 Pages Posted: 12 Sep 2019

See all articles by Robert Navratil

Robert Navratil

Charles University in Prague - Faculty of Mathematics and Physics

Jan Vecer

Charles University in Prague - Faculty of Mathematics and Physics

Date Written: September 4, 2019

Abstract

This paper presents a novel approach to model selection and model averaging based on economic theory. We study model prediction in the form of a distributional opinion about a random variable X. We show how to test this prediction against alternative views. Different model opinions can be traded on a hypothetical market that trades their differences. Using a utility maximization technique, we describe such a market for any general random variable X and any utility function U. We specify the optimal behavior of agents and the total market that aggregates all available opinions and show that a correct distributional opinion realizes profit in expectation against any other opinion, giving a novel technique for model selection. Analytical solutions are available for random variables from the exponential family. We determine the distribution corresponding to the aggregated view of all available opinions, giving a novel technique for model averaging.

Keywords: Utility Maximization, Explicit Demand Function, Equilibrium, Exponential Family, Model Selection, Model Averaging

JEL Classification: C52, C68, D58, G13

Suggested Citation

Navratil, Robert and Vecer, Jan, Utility Based Model Selection and Model Averaging (September 4, 2019). Available at SSRN: https://ssrn.com/abstract=3448046 or http://dx.doi.org/10.2139/ssrn.3448046

Robert Navratil

Charles University in Prague - Faculty of Mathematics and Physics ( email )

Sokolovska 83
Prague, 186 75
Czech Republic

Jan Vecer (Contact Author)

Charles University in Prague - Faculty of Mathematics and Physics ( email )

Sokolovska 83
Prague, 186 75
Czech Republic

Do you have negative results from your research you’d like to share?

Paper statistics

Downloads
149
Abstract Views
826
Rank
354,722
PlumX Metrics