Exact inference using variable integrating constant importance distributions

Charles J. Romeo, Teruo Nakatsuma

Research output: Contribution to journalArticle

Abstract

In this paper a new class of importance distributions that are based on normal and t-distributions are introduced. They are labeled variable integrating constant distributions or VIC distributions. The main innovation in this class of distributions is that the standard error associated with each prameter is expanded in an exponential polynomial function of the scaled distance from its mode, with a different polynomial used in each direction and for each parameter. Resulting variances differ at each point in the parameter space, thereby breaking the flexibility limitations that are inherent in normal and t-distributions allowing them to be closely matched to the posterior distribution whose moments are the goal of the inference exercise. Two illustrations are presented in which the performance of normal, t, normal-VIC, and t-VIC importance distributions are compared. Results show the VIC distributions to produce globally stable weights and substantial improvements in the numerical efficiency of importance sampling and independence Metropolis estimates.

Original languageEnglish
Pages (from-to)45-70
Number of pages26
JournalComputational Economics
Volume23
Issue number1
DOIs
Publication statusPublished - 2004

Fingerprint

Polynomials
Importance sampling
Innovation
Exact inference

Keywords

  • Importance sampling
  • Independence Metropolis
  • Variable integrating constant distributions

ASJC Scopus subject areas

  • Economics, Econometrics and Finance (miscellaneous)

Cite this

Exact inference using variable integrating constant importance distributions. / Romeo, Charles J.; Nakatsuma, Teruo.

In: Computational Economics, Vol. 23, No. 1, 2004, p. 45-70.

Research output: Contribution to journalArticle

@article{d806648d948143cf8cf9624248a6f0bc,
title = "Exact inference using variable integrating constant importance distributions",
abstract = "In this paper a new class of importance distributions that are based on normal and t-distributions are introduced. They are labeled variable integrating constant distributions or VIC distributions. The main innovation in this class of distributions is that the standard error associated with each prameter is expanded in an exponential polynomial function of the scaled distance from its mode, with a different polynomial used in each direction and for each parameter. Resulting variances differ at each point in the parameter space, thereby breaking the flexibility limitations that are inherent in normal and t-distributions allowing them to be closely matched to the posterior distribution whose moments are the goal of the inference exercise. Two illustrations are presented in which the performance of normal, t, normal-VIC, and t-VIC importance distributions are compared. Results show the VIC distributions to produce globally stable weights and substantial improvements in the numerical efficiency of importance sampling and independence Metropolis estimates.",
keywords = "Importance sampling, Independence Metropolis, Variable integrating constant distributions",
author = "Romeo, {Charles J.} and Teruo Nakatsuma",
year = "2004",
doi = "10.1023/B:CSEM.0000007188.61993.91",
language = "English",
volume = "23",
pages = "45--70",
journal = "Computer Science in Economics and Management",
issn = "0921-2736",
publisher = "Springer Netherlands",
number = "1",

}

TY - JOUR

T1 - Exact inference using variable integrating constant importance distributions

AU - Romeo, Charles J.

AU - Nakatsuma, Teruo

PY - 2004

Y1 - 2004

N2 - In this paper a new class of importance distributions that are based on normal and t-distributions are introduced. They are labeled variable integrating constant distributions or VIC distributions. The main innovation in this class of distributions is that the standard error associated with each prameter is expanded in an exponential polynomial function of the scaled distance from its mode, with a different polynomial used in each direction and for each parameter. Resulting variances differ at each point in the parameter space, thereby breaking the flexibility limitations that are inherent in normal and t-distributions allowing them to be closely matched to the posterior distribution whose moments are the goal of the inference exercise. Two illustrations are presented in which the performance of normal, t, normal-VIC, and t-VIC importance distributions are compared. Results show the VIC distributions to produce globally stable weights and substantial improvements in the numerical efficiency of importance sampling and independence Metropolis estimates.

AB - In this paper a new class of importance distributions that are based on normal and t-distributions are introduced. They are labeled variable integrating constant distributions or VIC distributions. The main innovation in this class of distributions is that the standard error associated with each prameter is expanded in an exponential polynomial function of the scaled distance from its mode, with a different polynomial used in each direction and for each parameter. Resulting variances differ at each point in the parameter space, thereby breaking the flexibility limitations that are inherent in normal and t-distributions allowing them to be closely matched to the posterior distribution whose moments are the goal of the inference exercise. Two illustrations are presented in which the performance of normal, t, normal-VIC, and t-VIC importance distributions are compared. Results show the VIC distributions to produce globally stable weights and substantial improvements in the numerical efficiency of importance sampling and independence Metropolis estimates.

KW - Importance sampling

KW - Independence Metropolis

KW - Variable integrating constant distributions

UR - http://www.scopus.com/inward/record.url?scp=26844548730&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=26844548730&partnerID=8YFLogxK

U2 - 10.1023/B:CSEM.0000007188.61993.91

DO - 10.1023/B:CSEM.0000007188.61993.91

M3 - Article

VL - 23

SP - 45

EP - 70

JO - Computer Science in Economics and Management

JF - Computer Science in Economics and Management

SN - 0921-2736

IS - 1

ER -