Decomposition of Some Well-Known Variance Reduction Techniques

Research output: Contribution to journalArticlepeer-review

20 Scopus citations

Abstract

Proliferation of techniques and lack of unifying structure have hindred both study and application of variance reduction. We view any variance reduction technique (VRT) as a transformation from one experiment to another, which leads to an exhaustive taxonomy of VRTs. In this paper, seven familiar VRTs are expressed in our taxonomy as compositions of elemental transformations from six basic classes. Our objective is to use the reader’s knowledge of these well-known techniques to illustrate the taxonomy.

Original languageEnglish (US)
Pages (from-to)183-209
Number of pages27
JournalJournal of Statistical Computation and Simulation
Volume23
Issue number3
DOIs
StatePublished - Jan 1 1986

Funding

The Office of Naval Research provided partial support via contract N00014-79-C-0832. We thank James R. Wilson for both general discussions and specific suggestions that have improved this paper and clarified our thoughts. tResearch partially supported by Ofice of Naval Research contract N00014-79-C-0832.

Keywords

  • Antithetic variates
  • Monte Carlo
  • common random numbers
  • conditional expectations
  • control variates
  • importance sampling
  • poststratified sampling
  • simulation
  • swindles
  • variance reduction

ASJC Scopus subject areas

  • Statistics and Probability
  • Modeling and Simulation
  • Statistics, Probability and Uncertainty
  • Applied Mathematics

Fingerprint

Dive into the research topics of 'Decomposition of Some Well-Known Variance Reduction Techniques'. Together they form a unique fingerprint.

Cite this