Orthogonal column latin hypercubes and their application in computer experiments

Research output: Contribution to journalArticle

268 Citations (Scopus)

Abstract

Latin hypercubes have been frequently used in conducting computer experiments. In this paper, a class of orthogonal Latin hypercubes that preserves orthogonality among columns is proposed. Applying an orthogonal Latin hypercube design to a computer experiment benefits the data analysis in two ways. First, it retains the orthogonality of traditional experimental designs. The estimates of linear effects of all factors are uncorrelated not only with each other, but also with the estimates of all quadratic effects and bilinear interactions. Second, it can facilitate nonparametric fitting procedures, because one can select good space-filling designs within the class of orthogonal Latin hypercubes according to selection criteria.

Original languageEnglish (US)
Pages (from-to)1430-1439
Number of pages10
JournalJournal of the American Statistical Association
Volume93
Issue number444
StatePublished - Dec 1998
Externally publishedYes

Fingerprint

Latin Hypercube
Computer Experiments
Orthogonality
Latin Hypercube Design
Orthogonal Design
Experimental design
Estimate
Data analysis
Interaction
Computer experiments
Class

Keywords

  • Computer model
  • Linear regression
  • Optimal design
  • Response surface design

ASJC Scopus subject areas

  • Mathematics(all)
  • Statistics and Probability

Cite this

Orthogonal column latin hypercubes and their application in computer experiments. / Ye, Qian K.

In: Journal of the American Statistical Association, Vol. 93, No. 444, 12.1998, p. 1430-1439.

Research output: Contribution to journalArticle

@article{0c73d0816c4f4a408c3b334016a5c13b,
title = "Orthogonal column latin hypercubes and their application in computer experiments",
abstract = "Latin hypercubes have been frequently used in conducting computer experiments. In this paper, a class of orthogonal Latin hypercubes that preserves orthogonality among columns is proposed. Applying an orthogonal Latin hypercube design to a computer experiment benefits the data analysis in two ways. First, it retains the orthogonality of traditional experimental designs. The estimates of linear effects of all factors are uncorrelated not only with each other, but also with the estimates of all quadratic effects and bilinear interactions. Second, it can facilitate nonparametric fitting procedures, because one can select good space-filling designs within the class of orthogonal Latin hypercubes according to selection criteria.",
keywords = "Computer model, Linear regression, Optimal design, Response surface design",
author = "Ye, {Qian K.}",
year = "1998",
month = "12",
language = "English (US)",
volume = "93",
pages = "1430--1439",
journal = "Journal of the American Statistical Association",
issn = "0162-1459",
publisher = "Taylor and Francis Ltd.",
number = "444",

}

TY - JOUR

T1 - Orthogonal column latin hypercubes and their application in computer experiments

AU - Ye, Qian K.

PY - 1998/12

Y1 - 1998/12

N2 - Latin hypercubes have been frequently used in conducting computer experiments. In this paper, a class of orthogonal Latin hypercubes that preserves orthogonality among columns is proposed. Applying an orthogonal Latin hypercube design to a computer experiment benefits the data analysis in two ways. First, it retains the orthogonality of traditional experimental designs. The estimates of linear effects of all factors are uncorrelated not only with each other, but also with the estimates of all quadratic effects and bilinear interactions. Second, it can facilitate nonparametric fitting procedures, because one can select good space-filling designs within the class of orthogonal Latin hypercubes according to selection criteria.

AB - Latin hypercubes have been frequently used in conducting computer experiments. In this paper, a class of orthogonal Latin hypercubes that preserves orthogonality among columns is proposed. Applying an orthogonal Latin hypercube design to a computer experiment benefits the data analysis in two ways. First, it retains the orthogonality of traditional experimental designs. The estimates of linear effects of all factors are uncorrelated not only with each other, but also with the estimates of all quadratic effects and bilinear interactions. Second, it can facilitate nonparametric fitting procedures, because one can select good space-filling designs within the class of orthogonal Latin hypercubes according to selection criteria.

KW - Computer model

KW - Linear regression

KW - Optimal design

KW - Response surface design

UR - http://www.scopus.com/inward/record.url?scp=0032286861&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0032286861&partnerID=8YFLogxK

M3 - Article

AN - SCOPUS:0032286861

VL - 93

SP - 1430

EP - 1439

JO - Journal of the American Statistical Association

JF - Journal of the American Statistical Association

SN - 0162-1459

IS - 444

ER -