Exact AD HOC inference from Weibull samples

John I. McCool

Research output: Contribution to journalArticle

1 Citation (Scopus)

Abstract

Exact inference based on type II censored Weibull samples depends on the availability of tabled values of percentage points of certain pivotal functions which must be found by Monte Carlo simulation. Applications have been limited by the comparatively small number of cases for which percentage points have been tabulated. The thesis of this paper is that computer hardware and software have now developed to the point where it is practical with the aid of a comparatively simple simulation program and a readily available spreadsheet or statistical package, for a user to determine the exact values needed in a number of inferential settings. It is shown how inferences on any Weibull percentile may be obtained once a single set of simulation results is found for any arbitrary percentile and the shape parameter. The same approach is shown to hold for inference about series systems of identical Weibull components and for computing the OC curve of a hypothesis test on a Weibull percentile. Finally, an example is given of the use of a spreadsheet to operate on the results of repeated sets of simulation runs to develop the values needed to conduct a multiple comparison analysis of a set of 13 Weibull distributed fracture data samples.

Original languageEnglish (US)
Pages (from-to)26-32
Number of pages7
JournalInternational Journal of Industrial Engineering : Theory Applications and Practice
Volume7
Issue number1
StatePublished - 2000

Fingerprint

Spreadsheets
Computer hardware
Availability
Monte Carlo simulation

All Science Journal Classification (ASJC) codes

  • Industrial and Manufacturing Engineering

Cite this

@article{56c50285173a4489b47f85594ce02960,
title = "Exact AD HOC inference from Weibull samples",
abstract = "Exact inference based on type II censored Weibull samples depends on the availability of tabled values of percentage points of certain pivotal functions which must be found by Monte Carlo simulation. Applications have been limited by the comparatively small number of cases for which percentage points have been tabulated. The thesis of this paper is that computer hardware and software have now developed to the point where it is practical with the aid of a comparatively simple simulation program and a readily available spreadsheet or statistical package, for a user to determine the exact values needed in a number of inferential settings. It is shown how inferences on any Weibull percentile may be obtained once a single set of simulation results is found for any arbitrary percentile and the shape parameter. The same approach is shown to hold for inference about series systems of identical Weibull components and for computing the OC curve of a hypothesis test on a Weibull percentile. Finally, an example is given of the use of a spreadsheet to operate on the results of repeated sets of simulation runs to develop the values needed to conduct a multiple comparison analysis of a set of 13 Weibull distributed fracture data samples.",
author = "McCool, {John I.}",
year = "2000",
language = "English (US)",
volume = "7",
pages = "26--32",
journal = "International Journal of Industrial Engineering : Theory Applications and Practice",
issn = "1072-4761",
publisher = "University of Cincinnati",
number = "1",

}

Exact AD HOC inference from Weibull samples. / McCool, John I.

In: International Journal of Industrial Engineering : Theory Applications and Practice, Vol. 7, No. 1, 2000, p. 26-32.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Exact AD HOC inference from Weibull samples

AU - McCool, John I.

PY - 2000

Y1 - 2000

N2 - Exact inference based on type II censored Weibull samples depends on the availability of tabled values of percentage points of certain pivotal functions which must be found by Monte Carlo simulation. Applications have been limited by the comparatively small number of cases for which percentage points have been tabulated. The thesis of this paper is that computer hardware and software have now developed to the point where it is practical with the aid of a comparatively simple simulation program and a readily available spreadsheet or statistical package, for a user to determine the exact values needed in a number of inferential settings. It is shown how inferences on any Weibull percentile may be obtained once a single set of simulation results is found for any arbitrary percentile and the shape parameter. The same approach is shown to hold for inference about series systems of identical Weibull components and for computing the OC curve of a hypothesis test on a Weibull percentile. Finally, an example is given of the use of a spreadsheet to operate on the results of repeated sets of simulation runs to develop the values needed to conduct a multiple comparison analysis of a set of 13 Weibull distributed fracture data samples.

AB - Exact inference based on type II censored Weibull samples depends on the availability of tabled values of percentage points of certain pivotal functions which must be found by Monte Carlo simulation. Applications have been limited by the comparatively small number of cases for which percentage points have been tabulated. The thesis of this paper is that computer hardware and software have now developed to the point where it is practical with the aid of a comparatively simple simulation program and a readily available spreadsheet or statistical package, for a user to determine the exact values needed in a number of inferential settings. It is shown how inferences on any Weibull percentile may be obtained once a single set of simulation results is found for any arbitrary percentile and the shape parameter. The same approach is shown to hold for inference about series systems of identical Weibull components and for computing the OC curve of a hypothesis test on a Weibull percentile. Finally, an example is given of the use of a spreadsheet to operate on the results of repeated sets of simulation runs to develop the values needed to conduct a multiple comparison analysis of a set of 13 Weibull distributed fracture data samples.

UR - http://www.scopus.com/inward/record.url?scp=0033719439&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0033719439&partnerID=8YFLogxK

M3 - Article

AN - SCOPUS:0033719439

VL - 7

SP - 26

EP - 32

JO - International Journal of Industrial Engineering : Theory Applications and Practice

JF - International Journal of Industrial Engineering : Theory Applications and Practice

SN - 1072-4761

IS - 1

ER -