Modern Industrial Statistics

STATISTICS IN PRACTICE

Series Advisors

Human and Biological Sciences

Stephen Senn

CRP-Sante,

́ Luxembourg

Earth and Environmental Sciences

Marian Scott

University of Glasgow, UK

Industry, Commerce and Finance

Wolfgang Jank

University of Maryland, USA

Founding Editor

Vic Barnett

Nottingham Trent University, UK

Statistics in Practice is an important international series of texts which provide detailed coverage of

statistical concepts, methods and worked case studies in specific fields of investigation and study.

With sound motivation and many worked practical examples, the books show in down-to-earth

terms how to select and use an appropriate range of statistical techniques in a particular practical

field within each title’s special topic area.

The books provide statistical support for professionals and research workers across a range of

employment fields and research environments. Subject areas covered include medicine and pharmaceutics; industry, finance and commerce; public services; the earth and environmental sciences,

and so on.

The books also provide support to students studying statistical courses applied to the above areas.

The demand for graduates to be equipped for the work environment has led to such courses becoming increasingly prevalent at universities and colleges.

It is our aim to present judiciously chosen and well-written workbooks to meet everyday practical

needs. Feedback of views from readers will be most valuable to monitor the success of this aim.

A complete list of titles in this series appears at the end of the volume.

Modern Industrial Statistics

with applications in R, MINITAB and JMP

Second Edition

RON S. KENETT

Chairman and CEO, the KPA Group, Israel

Research Professor, University of Turin, Italy, and

International Professor, NYU, Center for Risk Engineering, New York, USA

SHELEMYAHU ZACKS

Distinguished Professor,

Binghamton University, Binghamton, USA

With contributions from

DANIELE AMBERTI

Turin, Italy

This edition first published 2014

© 2014 John Wiley & Sons, Ltd

Registered office

John Wiley & Sons Ltd, The Atrium, Southern Gate, Chichester, West Sussex, PO19 8SQ, United Kingdom

For details of our global editorial offices, for customer services and for information about how to apply for permission to reuse the

copyright material in this book please see our website at www.wiley.com.

The right of the author to be identified as the author of this work has been asserted in accordance with the Copyright, Designs and

Patents Act 1988.

All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any

means, electronic, mechanical, photocopying, recording or otherwise, except as permitted by the UK Copyright, Designs and Patents

Act 1988, without the prior permission of the publisher.

Wiley also publishes its books in a variety of electronic formats. Some content that appears in print may not be available in electronic

books.

Designations used by companies to distinguish their products are often claimed as trademarks. All brand names and product names

used in this book are trade names, service marks, trademarks or registered trademarks of their respective owners. The publisher is not

associated with any product or vendor mentioned in this book.

Limit of Liability/Disclaimer of Warranty: While the publisher and author have used their best efforts in preparing this book, they make

no representations or warranties with respect to the accuracy or completeness of the contents of this book and specifically disclaim any

implied warranties of merchantability or fitness for a particular purpose. It is sold on the understanding that the publisher is not

engaged in rendering professional services and neither the publisher nor the author shall be liable for damages arising herefrom. If

professional advice or other expert assistance is required, the services of a competent professional should be sought.

Library of Congress Cataloging-in-Publication Data

Kenett, Ron.

Modern industrial statistics : with applications in R, MINITAB and JMP / Ron S. Kenett, Shelemyahu

Zacks. – Second edition.

pages cm

Includes bibliographical references and index.

ISBN 978-1-118-45606-4 (cloth)

1. Quality control–Statistical methods. 2. Reliability (Engineering)–Statistical methods. 3. R (Computer

program language) 4. Minitab. 5. JMP (Computer file) I. Zacks, Shelemyahu, 1932- II. Title.

TS156.K42 2014

658.5′ 62–dc23

2013031273

A catalogue record for this book is available from the British Library.

ISBN: 978-1-118-45606-4

Typeset in 9/11pt TimesLTStd by Laserwords Private Limited, Chennai, India

1

2014

To my wife Sima, our children and their children: Yonatan, Alma, Tomer,

Yadin, Aviv and Gili.

RSK

To my wife Hanna, our sons Yuval and David, and their families with love.

SZ

To my wife Nadia, and my family. With a special thought to my mother and

thank you to my father.

DA

Contents

Preface to Second Edition

Preface to First Edition

xv

xvii

Abbreviations

xix

PART I

1

The Role of Statistical Methods in Modern Industry and Services

1.1

1.2

1.3

1.4

1.5

1.6

1.7

1.8

1.9

2

The different functional areas in industry and services

The quality-productivity dilemma

Fire-fighting

Inspection of products

Process control

Quality by design

Information quality and practical statistical efficiency

Chapter highlights

Exercises

1

3

3

5

6

7

7

8

9

11

12

Analyzing Variability: Descriptive Statistics

13

2.1

2.2

2.3

2.4

13

17

18

19

19

23

26

28

32

32

33

34

34

36

38

38

2.5

2.6

2.7

2.8

3

PRINCIPLES OF STATISTICAL THINKING AND ANALYSIS

Random phenomena and the structure of observations

Accuracy and precision of measurements

The population and the sample

Descriptive analysis of sample values

2.4.1 Frequency distributions of discrete random variables

2.4.2 Frequency distributions of continuous random variables

2.4.3 Statistics of the ordered sample

2.4.4 Statistics of location and dispersion

Prediction intervals

Additional techniques of exploratory data analysis

2.6.1 Box and whiskers plot

2.6.2 Quantile plots

2.6.3 Stem-and-leaf diagrams

2.6.4 Robust statistics for location and dispersion

Chapter highlights

Exercises

Probability Models and Distribution Functions

41

3.1

41

41

42

44

Basic probability

3.1.1 Events and sample spaces: Formal presentation of random measurements

3.1.2 Basic rules of operations with events: Unions, intersections

3.1.3 Probabilities of events

viii

4

Contents

3.1.4 Probability functions for random sampling

3.1.5 Conditional probabilities and independence of events

3.1.6 Bayes formula and its application

3.2 Random variables and their distributions

3.2.1 Discrete and continuous distributions

3.2.2 Expected values and moments of distributions

3.2.3 The standard deviation, quantiles, measures of skewness and kurtosis

3.2.4 Moment generating functions

3.3 Families of discrete distribution

3.3.1 The binomial distribution

3.3.2 The hypergeometric distribution

3.3.3 The Poisson distribution

3.3.4 The geometric and negative binomial distributions

3.4 Continuous distributions

3.4.1 The uniform distribution on the interval (a, b), a < b

3.4.2 The normal and log-normal distributions

3.4.3 The exponential distribution

3.4.4 The gamma and Weibull distributions

3.4.5 The Beta distributions

3.5 Joint, marginal and conditional distributions

3.5.1 Joint and marginal distributions

3.5.2 Covariance and correlation

3.5.3 Conditional distributions

3.6 Some multivariate distributions

3.6.1 The multinomial distribution

3.6.2 The multi-hypergeometric distribution

3.6.3 The bivariate normal distribution

3.7 Distribution of order statistics

3.8 Linear combinations of random variables

3.9 Large sample approximations

3.9.1 The law of large numbers

3.9.2 The Central Limit Theorem

3.9.3 Some normal approximations

3.10 Additional distributions of statistics of normal samples

3.10.1 Distribution of the sample variance

3.10.2 The “Student” t-statistic

3.10.3 Distribution of the variance ratio

3.11 Chapter highlights

3.12 Exercises

46

47

49

51

51

55

57

59

60

60

62

65

67

69

69

70

75

77

80

82

82

84

86

88

88

89

90

92

94

98

98

99

99

101

101

102

102

104

105

Statistical Inference and Bootstrapping

113

4.1

4.2

113

114

115

116

118

120

120

122

128

4.3

4.4

Sampling characteristics of estimators

Some methods of point estimation

4.2.1 Moment equation estimators

4.2.2 The method of least squares

4.2.3 Maximum likelihood estimators

Comparison of sample estimates

4.3.1 Basic concepts

4.3.2 Some common one-sample tests of hypotheses

Confidence intervals

Contents ix

4.5

4.6

4.7

4.8

4.9

4.10

4.11

4.12

4.13

4.14

4.15

4.16

5

4.4.1 Confidence intervals for 𝜇; 𝜎 known

4.4.2 Confidence intervals for 𝜇; 𝜎 unknown

4.4.3 Confidence intervals for 𝜎 2

4.4.4 Confidence intervals for p

Tolerance intervals

4.5.1 Tolerance intervals for the normal distributions

Testing for normality with probability plots

Tests of goodness of fit

4.7.1 The chi-square test (large samples)

4.7.2 The Kolmogorov-Smirnov test

Bayesian decision procedures

4.8.1 Prior and posterior distributions

4.8.2 Bayesian testing and estimation

4.8.3 Credibility intervals for real parameters

Random sampling from reference distributions

Bootstrap sampling

4.10.1 The bootstrap method

4.10.2 Examining the bootstrap method

4.10.3 Harnessing the bootstrap method

Bootstrap testing of hypotheses

4.11.1 Bootstrap testing and confidence intervals for the mean

4.11.2 Studentized test for the mean

4.11.3 Studentized test for the difference of two means

4.11.4 Bootstrap tests and confidence intervals for the variance

4.11.5 Comparing statistics of several samples

Bootstrap tolerance intervals

4.12.1 Bootstrap tolerance intervals for Bernoulli samples

4.12.2 Tolerance interval for continuous variables

4.12.3 Distribution-free tolerance intervals

Non-parametric tests

4.13.1 The sign test

4.13.2 The randomization test

4.13.3 The Wilcoxon Signed Rank test

Description of MINITAB macros (available for download from Appendix VI of the book website)

Chapter highlights

Exercises

129

130

130

130

132

132

134

137

137

139

140

141

144

147

148

150

150

151

152

152

153

153

155

157

158

161

161

163

164

165

165

166

168

170

170

171

Variability in Several Dimensions and Regression Models

177

5.1

177

177

179

181

182

185

185

185

187

192

194

198

5.2

5.3

5.4

5.5

Graphical display and analysis

5.1.1 Scatterplots

5.1.2 Multiple boxplots

Frequency distributions in several dimensions

5.2.1 Bivariate joint frequency distributions

5.2.2 Conditional distributions

Correlation and regression analysis

5.3.1 Covariances and correlations

5.3.2 Fitting simple regression lines to data

Multiple regression

5.4.1 Regression on two variables

Partial regression and correlation

x

Contents

5.6

5.7

5.8

5.9

5.10

5.11

5.12

5.13

5.14

5.15

5.16

Multiple linear regression

Partial F-tests and the sequential SS

Model construction: Step-wise regression

Regression diagnostics

Quantal response analysis: Logistic regression

The analysis of variance: The comparison of means

5.11.1 The statistical model

5.11.2 The one-way analysis of variance (ANOVA)

Simultaneous confidence intervals: Multiple comparisons

Contingency tables

5.13.1 The structure of contingency tables

5.13.2 Indices of association for contingency tables

Categorical data analysis

5.14.1 Comparison of binomial experiments

Chapter highlights

Exercises

200

204

206

209

211

213

213

214

216

220

220

223

227

227

229

230

PART II ACCEPTANCE SAMPLING

235

6

Sampling for Estimation of Finite Population Quantities

237

6.1

237

237

238

239

241

242

244

248

249

252

255

256

6.2

6.3

6.4

6.5

6.6

6.7

7

Sampling and the estimation problem

6.1.1 Basic definitions

6.1.2 Drawing a random sample from a finite population

6.1.3 Sample estimates of population quantities and their sampling distribution

Estimation with simple random samples

6.2.1 Properties of X n and Sn2 under RSWR

6.2.2 Properties of X n and Sn2 under RSWOR

Estimating the mean with stratified RSWOR

Proportional and optimal allocation

Prediction models with known covariates

Chapter highlights

Exercises

Sampling Plans for Product Inspection

258

7.1

7.2

7.3

7.4

7.5

7.6

7.7

7.8

7.9

258

259

262

264

267

270

272

274

276

276

278

279

280

281

7.10

7.11

7.12

7.13

General discussion

Single-stage sampling plans for attributes

Approximate determination of the sampling plan

Double-sampling plans for attributes

Sequential sampling

Acceptance sampling plans for variables

Rectifying inspection of lots

National and international standards

Skip-lot sampling plans for attributes

7.9.1 The ISO 2859 skip-lot sampling procedures

The Deming inspection criterion

Published tables for acceptance sampling

Chapter highlights

Exercises

Contents xi

PART III STATISTICAL PROCESS CONTROL

283

8

Basic Tools and Principles of Process Control

285

8.1

8.2

8.3

8.4

8.5

8.6

8.7

285

294

298

300

302

305

308

309

311

316

316

8.8

8.9

9

Advanced Methods of Statistical Process Control

319

9.1

319

319

321

323

324

325

328

328

328

330

330

333

335

339

342

346

347

348

350

351

354

356

357

9.2

9.3

9.4

9.5

9.6

9.7

9.8

9.9

10

Basic concepts of statistical process control

Driving a process with control charts

Setting up a control chart: Process capability studies

Process capability indices

Seven tools for process control and process improvement

Statistical analysis of Pareto charts

The Shewhart control charts

8.7.1 Control charts for attributes

8.7.2 Control charts for variables

Chapter highlights

Exercises

Tests of randomness

9.1.1 Testing the number of runs

9.1.2 Runs above and below a specified level

9.1.3 Runs up and down

9.1.4 Testing the length of runs up and down

Modified Shewhart control charts for X

The size and frequency of sampling for Shewhart control charts

9.3.1 The economic design for X-charts

9.3.2 Increasing the sensitivity of p-charts

Cumulative sum control charts

9.4.1 Upper Page’s scheme

9.4.2 Some theoretical background

9.4.3 Lower and two-sided Page’s scheme

9.4.4 Average run length, probability of false alarm and conditional expected delay

Bayesian detection

Process tracking

9.6.1 The EWMA procedure

9.6.2 The BECM procedure

9.6.3 The Kalman filter

9.6.4 Hoadley’s QMP

Automatic process control

Chapter highlights

Exercises

Multivariate Statistical Process Control

361

10.1

10.2

10.3

10.4

361

365

367

370

370

370

371

372

Introduction

A review of multivariate data analysis

Multivariate process capability indices

Advanced applications of multivariate control charts

10.4.1 Multivariate control charts scenarios

10.4.2 Internally derived targets

10.4.3 Using an external reference sample

10.4.4 Externally assigned targets

xii

Contents

10.4.5 Measurement units considered as batches

10.4.6 Variable decomposition and monitoring indices

10.5 Multivariate tolerance specifications

10.6 Chapter highlights

10.7 Exercises

373

373

374

376

377

PART IV DESIGN AND ANALYSIS OF EXPERIMENTS

379

11

12

Classical Design and Analysis of Experiments

381

11.1

11.2

11.3

11.4

Basic steps and guiding principles

Blocking and randomization

Additive and non-additive linear models

The analysis of randomized complete block designs

11.4.1 Several blocks, two treatments per block: Paired comparison

11.4.2 Several blocks, t treatments per block

11.5 Balanced incomplete block designs

11.6 Latin square design

11.7 Full factorial experiments

11.7.1 The structure of factorial experiments

11.7.2 The ANOVA for full factorial designs

11.7.3 Estimating main effects and interactions

11.7.4 2m factorial designs

11.7.5 3m factorial designs

11.8 Blocking and fractional replications of 2m factorial designs

11.9 Exploration of response surfaces

11.9.1 Second order designs

11.9.2 Some specific second order designs

11.9.3 Approaching the region of the optimal yield

11.9.4 Canonical representation

11.10 Chapter highlights

11.11 Exercises

381

385

385

387

387

391

394

397

402

402

402

408

409

417

425

430

431

433

438

440

441

442

Quality by Design

446

12.1 Off-line quality control, parameter design and the Taguchi method

12.1.1 Product and process optimization using loss functions

12.1.2 Major stages in product and process design

12.1.3 Design parameters and noise factors

12.1.4 Parameter design experiments

12.1.5 Performance statistics

12.2 The effects of non-linearity

12.3 Taguchi’s designs

12.4 Quality by design in the pharmaceutical industry

12.4.1 Introduction to quality by design

12.4.2 A quality by design case study – the full factorial design

12.4.3 A quality by design case study – the profiler and desirability function

12.4.4 A quality by design case study – the design space

12.5 Tolerance designs

12.6 More case studies

12.6.1 The Quinlan experiment at Flex Products, Inc.

12.6.2 Computer response time optimization

447

447

448

449

450

452

452

456

458

458

459

462

462

462

467

467

469

Contents xiii

13

12.7 Chapter highlights

12.8 Exercises

474

474

Computer Experiments

477

13.1

13.2

13.3

13.4

13.5

13.6

13.7

477

481

483

488

491

492

492

Introduction to computer experiments

Designing computer experiments

Analyzing computer experiments

Stochastic emulators

Integrating physical and computer experiments

Chapter highlights

Exercises

PART V RELIABILITY AND SURVIVAL ANALYSIS

495

14

15

Reliability Analysis

497

14.1 Basic notions

14.1.1 Time categories

14.1.2 Reliability and related functions

14.2 System reliability

14.3 Availability of repairable systems

14.4 Types of observations on TTF

14.5 Graphical analysis of life data

14.6 Non-parametric estimation of reliability

14.7 Estimation of life characteristics

14.7.1 Maximum likelihood estimators for exponential TTF distribution

14.7.2 Maximum likelihood estimation of the Weibull parameters

14.8 Reliability demonstration

14.8.1 Binomial testing

14.8.2 Exponential distributions

14.9 Accelerated life testing

14.9.1 The Arrhenius temperature model

14.9.2 Other models

14.10 Burn-in procedures

14.11 Chapter highlights

14.12 Exercises

498

498

499

500

503

509

510

513

514

514

518

520

520

521

528

528

529

529

530

531

Bayesian Reliability Estimation and Prediction

534

15.1 Prior and posterior distributions

15.2 Loss functions and Bayes estimators

15.2.1 Distribution-free Bayes estimator of reliability

15.2.2 Bayes estimator of reliability for exponential life distributions

15.3 Bayesian credibility and prediction intervals

15.3.1 Distribution-free reliability estimation

15.3.2 Exponential reliability estimation

15.3.3 Prediction intervals

15.4 Credibility intervals for the asymptotic availability of repairable systems: The exponential case

15.5 Empirical Bayes method

15.6 Chapter highlights

15.7 Exercises

534

537

538

538

539

539

540

540

542

543

545

545

xiv

Contents

List of R Packages

References and Further Reading

547

549

Author Index

Subject Index

555

557

Also available on book’s website: www.wiley.com/go/modern_industrial_statistics

Appendix I: An Introduction to R by Stefano Iacus

Appendix II: Basic MINITAB Commands and a Review of Matrix Algebra for Statistics

Appendix III: mistat Manual (mistat.pdf) and List of R Scripts, by Chapter (R_scripts.zip)

Appendix IV: Source Version of mistat Package (mistat_1.0.tar.gz), also available on the

Comprehensive R Archive Network (CRAN) Website.

Appendix V: Data Sets as csv Files

Appendix VI: MINITAB Macros

Appendix VII: JMP Scripts by Ian Cox

Appendix VIII: Solution Manual

Preface to Second Edition

This book is about modern industrial statistics and it applications using R, MINITAB and JMP. It is an expanded second

edition of a book entitled Modern Industrial Statistics: Design and Control of Quality and Reliability, Wadsworth Duxbury

Publishing, 1998, Spanish edition: Estadistica Industrial Moderna: Diseño y Control de Calidad y la Confiabilidad,

Thomson International, 2000. Abbreviated edition: Modern Statistics: A Computer-based Approach, Thomson Learning,

2001. Chinese edition: China Statistics Press, 2003 and Softcover edition, Brooks-Cole, 2004.

The pedagogical structure of the book combines a practical approach with theoretical foundations and computer support. It is intended for students and instructors who have an interest in studying modern methods by combining these

three elements. The first edition referred to S-Plus, MINITAB and compiled QuickBasic code. In this second edition we

provide examples and procedures in the now popular R language and also refer to MINITAB and JMP. Each of these

three computer platforms carries unique advantages. Focusing on only one or two of these is also possible. Exercises are

provided at the end of each chapter in order to provide more opportunities to learn and test your knowledge.

R is an open source programming language and software environment for statistical computing and graphics based

on the S programming language created by John Chambers while at Bell Labs in 1976. It is now developed by the R

Development Core Team, of which Chambers is a member. MINITAB is a statistics package originally developed at

the Pennsylvania State University by Barbara Ryan, Thomas Ryan, Jr., and Brian Joiner in 1972. MINITAB began as a

light version of OMNITAB, a statistical analysis program developed at the National Bureau of Standards now called the

National Institute of Standards and Technology (NIST). JMP was originally written in 1989 by John Sall and others to

perform simple and complex statistical analyses by dynamically linking statistics with graphics to interactively explore,

understand, and visualize data. JMP stands for John’s Macintosh Project and it is a division of SAS Institute Inc.

A clear advantage of R is that it is free open source software. It requires, however, knowledge of command language

programming. To help the reader, we developed a special comprehensive R application called mistat that includes all the

R programs used in the book. MINITAB is a popular statistical software application providing extensive collaboration and

reporting capabilities. JMP, a product of the SAS company, is also very popular and carries advanced scripting features and

high level visualization components. Both R and JMP have fully compatible versions for Mac OS. We do not aim to teach

programming in R or using MINITAB or JMP. We also do not cover all the options and features available in MINITAB

and JMP. Our aim is to expose students, researchers and practitioners of modern industrial statistics to examples of what

can be done with these software platforms and encourage the exploration of additional options in MINITAB and JMP.

Eventually, availability and convenience determine what software is used in specific cirmcumstances. We provide here

an opportunity to learn and get acquainted with three leading modern industrial statistics software platforms. A specially

prepared appendix, downloadable from the book website, provides an introduction to R. Also available for download

are the R scripts we refer to, organized by chapter. Installations of JMP and MINITAB include effective tutorials with

introductory material. Such tutorials have not been replicated in this text. To take full advantage of this book you need

to be interested in industrial statistics, have a proper mathematical background and be willing to learn by working out

problems with software applications. The five parts of the book can be studied in various combinations with Part I used

as a foundation. The book can therefore be used in workshops or courses on Acceptance Sampling, Statistical Process

Control, Design of Experiments and Reliability.

The three software platforms we refer to provide several simulation options. We believe that teaching modern industrial

statistics, with simulations, provides the right context for gaining sound hands-on experience. We aim at the middle road

target, between theoretical treatment and a cookbook approach. To achieve this, we provide over 40 data sets representing

real-life case studies which are typical of what one finds while performing statistical work in business and industry.

Figures in the book have been produced with R and in MINITAB and JMP as explicitly stated. In this second edition we

include contributions by Dr. Daniele Amberti who developed the R mistat package and provided many inputs to the

text. His work was supported by i4C (www.i4CAnalytics.com) and we are very grateful for it. Another change in this

xvi

Preface to Second Edition

second edition is that it is published by John Wiley and Sons and we would like to thank Heather Kay and Richard Davies

for their professional support and encouragements throughout this effort. We also acknowledge the contribution of Ian

Cox from JMP who developed for us the simulation code running the piston simulator and the bootstrapping analysis.

Thanks are due to the Genova Hotel in front of Porta Nuova in Turin where we spend countless hours updating the text

using push and pull of LaTeX files in Git from within RStudio. In fact, the whole book, with its R code and data sets has

been fully compiled in its present form to create an example of what reproducible research is all about. Gerry Hahn, Murat

Testik, Moshe Pollak, Gejza Dohnal, Neil Ullman, Moshe Miller, David Steinberg, Marcello Fidaleo, Inbal Yahav and Ian

Cox provided feedback on early drafts of the new chapters included in this second expanded edition of the original 1998

book–we thank them for their insightful comments. Finally, we would like to acknowledge the help of Marco Giuliano

who translated most of TeX files from the 1998 edition to LaTeX and of Marge Pratt who helped produce the final LaTeX

version of our work.

The book is accompanied by a dedicated website where all software and data files used are available to download.

The book website URL is www.wiley.com/go/modern_industrial_statistics.

The site contains: 1) all the R code included in the book which is also available on the R CRAN website as the mistat

package (folder scripts), 2) a source version of the mistat package for R (mistat_1.0.tar.gz), 3) all data sets as csv files

(csvFiles.zip and folder csvFiles in package mistat), 4) the MINITAB macros and JMP add-ins used in the text, 5)

an introduction to R prepared by Professor Stefano Iacus from the University of Milan and 6) solutions to some of the

exercises. Specifically, the book web site includes eight appendices: Appendix I - Introduction to R, by Stefano Iacus,

Appendix II - Basic MINITAB commands and a review of matrix algebra for Statistics, Appendix III - R code included

in the book, Appendix IV - Source version of package mistat, Appendix V - Data sets as csv files, Appendix VI MINITAB macros, Appendix VII - JMP scripts, by Ian Cox and Appendix VIII - Solution manual.

Special thanks are due to Professor Iacus for his generous contribution. If you are not familiar with R, we recommend

you look at this introduction specially prepared by one of the most important core developers of R. The material on the

book website should be considered part of the book. We obviously look forward to feedback, comments and suggestions

from students, teachers, researchers and practitioners and hope the book will help these different target groups achieve

concrete and significant impact with the tools and methods of industrial statistics.

Ron S. Kenett

Raanana, Israel and Turin, Italy

ron@kpa-group.com

Shelemyahu Zacks

Binghamton, New York, USA

shelly@math.binghamton.edu

Preface to First Edition

Modern Industrial Statistics provides the tools for those who drive to achieve perfection in industrial processes. Learn the

concepts and methods contained in this book and you will understand what it takes to measure and improve world-class

products and services.

The need for constant improvement of industrial processes, in order to achieve high quality, reliability, productivity

and profitability, is well recognized. Furthermore management techniques, such as total quality management or business

process reengineering, are insufficient in themselves to achieve the goal without the strong backing of specially tailored

statistical procedures, as stated by Robert Galvin in the Foreword.

Statistical procedures, designed for solving industrial problems, are called Industrial Statistics. Our objective in writing

this book was to provide statistics and engineering students, as well as practitioners, the concepts, applications, and

practice of basic and advanced industrial statistical methods, which are designed for the control and improvement of

quality and reliability.

The idea of writing a text on industrial statistics developed after several years of collaboration in industrial consulting, teaching workshops and seminars, and courses at our universities. We felt that no existing text served our needs in

both content and approach so we decided to develop our notes into a text. Our aim was to make the text modern and

comprehensive in terms of the techniques covered, lucid in its presentation, and practical with regard to implementation.

Ron S. Kenett

Shelemyahu Zacks

1998

Abbreviations

ANOVA

ANSI

AOQ

AOQL

AQL

ARL

ASN

ASQ

ATI

BECM

BIBD

BP

CAD

CADD

CAM

CBD

c.d.f.

CED

cGMP

CIM

CLT

CNC

CPA

CQA

CUSUM

DFIT

DLM

DoE

EBD

EWMA

FPM

GRR

HPD

i.i.d.

InfoQ

IQR

ISC

KS

LCL

LLN

LSE

LQL

LSL

Analysis of Variance

American National Standard Institute

Average Outgoing Quality

Average Outgoing Quality Limit

Acceptable Quality Level

Average Run Length

Average Sample Number

American Society for Quality

Average Total Inspection

Bayes Estimation of the Current Mean

Balanced Incomplete Block Design

Bootstrap Population

Computer Aided Design

Computer Aided Drawing and Drafting

Computer Aided Manufacturing

Complete Block Design

cumulative distribution function

Conditional Expected Delay

current good manufacturing practices

Computer Integrated Manufacturing

Central Limit Theorem

Computerized Numerically Controlled

Circuit Pack Assemblies

Critical Quality Attribute

Cumulative Sum

Fits distance

Dynamic Linear Model

Design of Experiments

Empirical Bootstrap Distribution

Exponentially Weighted Moving Average

Failures Per Million

Gage Repeatability and Reproducibility

Highest Posterior Density

independent and identically distributed

Information Quality

Inter Quartile Range

Short Circuit Current of Solar Cells (in Ampere)

Kolmogorov-Smirnov

Lower Control Limit

Law of Large Numbers

Least Squares Estimators

Limiting Quality Level

Lower Specification Limit

xx

Abbreviations

LTPD

LWL

m.g.f.

MLE

MSD

MTBF

MTTF

OC

p.d.f.

PERT

PFA

PL

PPM

PSE

QbD

QMP

QQ-Plot

RCBD

RMSE

RSWOR

RSWR

SE

SL

SLOC

SLSP

SPC

SPRT

SR

SSE

SSR

SST

STD

TTC

TTF

TTR

TTT

UCL

USL

UWL

WSP

Lot Tolerance Percent Defective

Lower Warning Limit

moment generating function

Maximum Likelihood Estimator

Mean Squared Deviation

Mean Time Between Failures

Mean Time To Failure

Operating Characteristic

probability density function

Project Evaluation and Review Technique

Probability of False Alarm

Product Limit Estimator

Defects in Parts Per Million

Practical Statistical Efficiency

Quality by Design

Quality Measurement Plan

Quantile vs. Quantile Plot

Randomized Complete Block Design

Root Mean Squared Error

Random Sample Without Replacement

Random Sample With Replacement

Standard Error

Skip Lot

Source Lines of Code

Skip Lot Sampling Plans

Statistical Process Control

Sequential Probability Ratio Test

Shiryayev-Roberts

Sum of Squares of Residuals

Sum of Squares Around the Regression Model

Total Sum of Squares

Standard Deviation

Time Till Censoring

Time Till Failure

Time Till Repair

Total Time on Test

Upper Control Limit

Upper Specification Limit

Upper Warning Limit

Wave Soldering Process

Part I

Principles of Statistical Thinking and Analysis

Part I is an introduction to the role of statistics in modern industry and service organizations, and to statistical thinking in

general. Typical industrial problems are described and basic statistical concepts and tools are presented through case studies and computer simulations. To help focus on data analysis and interpretation of results we refer, throughout the book, to

three leading software platforms for statistical analysis: R, MINITAB and JMP. R is an open source programming language

with more than 4300 application packages available at the Comprehensive R Archive Network (http://cran.r-project.org/).

MINITAB and JMP are statistical packages widely used in business and industry (www.minitab.com, www.jmp.com) with

30 days free fully functional downloads. The R code is integrated in the text and packaged in the mistat R application

that can be downloaded from CRAN or the book’s website www.wiley.com/go/modern_industrial_statistics. For ease of

use the mistat R applications are also organized by chapter in a set of scripts. The chapter scripts include all the R

examples used in the specific chapter.

Chapter 1 is an overview of the role of statistical methods in industry and offers a classification of statistical methods in

the context of various management approaches. We call this classification the Quality Ladder and use it to organize the

methods of industrial statistics covered in Parts II–V. The chapter also introduces the reader to the concept of practical

statistical efficiency (PSE) and information quality (InfoQ) that are used to assess the impact of work performed on a

given data set with statistical tools and the quality of knowledge generated by statistical analysis.

Chapter 2 presents basic concepts and tools for describing variability. It emphasizes graphical techniques to explore

and summarize variability in observations. The chapter introduces the reader to R and provides examples in MINITAB

and JMP. The examples demonstrate capabilities but it was not our intention to present introductory material on how to

use these software applications. Some help in this can be found in the book downloadable appendices.

Chapter 3 is an introduction to probability models including a comprehensive treatment of statistical distributions that

have applicability to industrial statistics. The chapter provides a reference to fundamental results and basic principles used

in later chapters.

Chapter 4 is dedicated to statistical inference and bootstrapping. Bootstrapping is introduced with examples in R,

MINITAB and JMP. With this approach, statistical procedures used in making inference from a sample to a population

are handled by computer-intensive procedures without the traditional need to validate mathematical assumptions and

models.

Chapter 5 deals with variability in several dimensions and regression models. It begins with graphical techniques that

handle observations taken simultaneously on several variables. These techniques are now widely available in software

applications such as those used throughout the book. The chapter covers linear and multiple regression models including

diagnostics and prediction intervals. Categorical data and multi-dimensional contingency tables are also analyzed.

Modern Industrial Statistics: with applications in R, MINITAB and JMP, Second Edition. Ron S. Kenett and Shelemyahu Zacks.

© 2014 John Wiley & Sons, Ltd. Published 2014 by John Wiley & Sons, Ltd.

Companion website: www.wiley.com/go/modern_industrial_statistics

1

The Role of Statistical Methods in

Modern Industry and Services

1.1

The different functional areas in industry and services

Industrial statistics has played a key role in the creation of competitiveness in a wide range of organizations in the industrial sector, in services, health care, government and educational systems. The tools and concepts of industrial statistics

have to be viewed in the context of their applications. These applications are greatly affected by management style and

organizational culture. We begin by describing key aspects of the industrial setting in order to lay out the foundations for

the book.

Industrial organizations typically include units dedicated to product development, manufacturing, marketing, finance,

human resources, purchasing, sales, quality assurance and after-sales support. Industrial statistics is used to resolve problems in each one of these functional units. Marketing personnel determine customer requirements and measure levels

of customer satisfaction using surveys and focus groups. Sales are responsible for providing forecasts to purchasing and

manufacturing. Purchasing specialists analyze world trends in quality and prices of raw materials so that they can optimize

costs and delivery time. Budgets are prepared by the finance department using forecasts that are validated periodically.

Accounting experts rely on auditing and sampling methods to ascertain inventory levels and integrity of databases. Human

resources personnel track data on absenteeism, turnover, overtime and training needs. They also conduct employee surveys and deploy performance appraisal systems. The quality departments commonly perform audits and quality tests,

to determine and ensure the quality and reliability of products and services. Research and development engineers perform experiments to solve problems and improve products and processes. Finally, manufacturing personnel and process

engineers design process controls for production operations using control charts and automation.

These are only a few examples of problem areas where the tools of industrial statistics are used within modern industrial

and service organizations. In order to provide more specific examples we first take a closer look at a variety of industries.

Later we discuss examples from these types of industries.

There are basically three types of production systems: (1) continuous flow production; (2) job shops; and (3) discrete

mass production. Examples of continuous flow production include steel, glass and paper making, thermal power

generation and chemical transformations. Such processes typically involve expensive equipment that is very large

in size, operates around the clock and requires very rigid manufacturing steps. Continuous flow industries are both

capital-intensive and highly dependent on the quality of the purchased raw materials. Rapid customizing of products in a

continuous flow process is virtually impossible and new products are introduced using complex scale-up procedures.

Modern Industrial Statistics: with applications in R, MINITAB and JMP, Second Edition. Ron S. Kenett and Shelemyahu Zacks.

© 2014 John Wiley & Sons, Ltd. Published 2014 by John Wiley & Sons, Ltd.

Companion website: www.wiley.com/go/modern_industrial_statistics

## Tư duy hệ thống

## Tư duy hệ thống và đổi mới tư duy

## Trình bày các khái niệm liên quan đến hệ thống khắc phục lỗi và xây dựng một hệ thống khắc phục lỗi bằng ngôn ngữ lập trình Java

## Bài 1: Khái niệm về tư pháp quốc tế và nguồn của tư pháp quốc tế

## Tư duy hệ thống - Nguồn sức sống mới của đổi mới tư duy - Phan Đình Diệu

## Tư duy hệ thống cơ bản

## Bài giảng TU DUY HE THONG

## Tài liệu Bài thuyết trình: Khái niệm Tín Hiệu và Hệ Thống doc

## Tài liệu Tư duy hệ thống cơ bản pdf

## Tài liệu Tư duy hệ thống (phần 1) pptx

Tài liệu liên quan