TMA4255 Experimental design and applied statistical methods

Fall 2008

 

Messages

17.10.08 Exam is on Friday December 5, 2008 at 09:00, durance 4 hours.

 

02.10.08 Updated information on the obligatory project is now found here.

 

30.09.08 An updated version of the information in English about the obligatory exercise can now be downloaded under “Exercises” below.

 

25.09.08 The obligatory exercise will start at the end of Week 40 (October 3), and deadline for submitting the report Is November 7 at 12 Noon. There will be no new exercises in Weeks 41 and 42 (October 6 and 13), instead you may consult the teaching assistant in the regular exercise hours these weeks. There will also be scheduled additional times in these weeks for consulting with teaching assistant and lecturer.

11.09.08 There will be NO LECTURE Tuesday, September 16.

29.08.08 Some earlier exams w/solution are given under “Exam” below. Unfortunately all this is in Norwegian, except the May 2007 exam.

22.08.08 Here is the link to the exercise website. Note that the time for exercises has been changed to Mondays 16.15-17, first time August 25 in Vegas computer lab (see exercise website for details).

19.08.08 There is now just one homepage of this course, which is in ENGLISH (the Norwegian one is hence no longer available).

19.08.08 Date of exam will be given after September 15 (after final registration for courses).

19.08.08 First lecture. Decided to give all lectures in English. No changes made in the weekly lecture hours. Next lecture is Friday August 22 at 8.15 in R3.

About the course

The course is intended for students who want to increase their knowledge in statistics beyond the basic course TMA4240/45. Central topics are how to plan and perform an experiment in order to collect high quality data and how to analysis these data such that information as a basis for conclusions can be achieved. The content is: Experimental design methodology, 2k-designs and fractions thereof, Graphical techniques, Approximation of expectation and variance, Testing of hypothesis, Simple and multiple linear regression, Residual analysis and variable selection, Transformations, Analysis of variance, Process control, Contingency tables and Non-parametric statistics. All methods that are in extensive use in industry and technology in order to improve quality and productivity. Analysing the data will be done mainly by use of a statistical software called MINITAB. Necessary background for the course is TMA4240/TMA4245 Statistcs or similar background. The course puts less emphasize on theory than TMA4260 Industrial statistics.

  • Credit: 7.5 stp.
  • Duration: 1 semester (fall)
  • Lecturing: 4 hours a week
  • Exercises: 1 hour a week
  • Special activity: A compulsary project counts 20 % for the exam.
  • Exam: Written, 4 hours.

 

Lecturer

Professor Bo Lindqvist, Room 1129, Sentralbygg II, NTNU Gløshaugen
Telephone: 73 59 35 32
Email:
bo@math.ntnu.no
Office hours: Fridays 13:15 – 15:00

Teaching assistant

Stipendiat Inge Myrseth, Room 1026, Sentralbygg II.

Telephone: 73 59 04 84
Email:
inge.myrseth@math.ntnu.no

Office hours: Thursdays 16:15 – 17:00

Reference group

To be determined

Teaching material

Main book: Walpole, Myers, Myers and Ye: "Probability and Statistics for Engineers and Scientists". Seventh or Eighth Edition. Pearson/Prentice Hall.

Handout (copy): Høyland, A (1988): Sannsynlighetsregning, Tapir. Kap. 5.2.8 Approksimasjon av forventning og varians til funksjoner av stokastiske variabler.

Handout (copy): From Box, Hunter and Hunter (1978): "Statistics for Experimenters. An introduction to Design, Data Analysis and Model Building". Wiley. Selected parts from chapters 10 and 12.

Tables: "Statistiske tabeller og formler". Tapir.

MINITAB: Some information is found on http://www.ntnu.no/adm/it/brukerstotte/programvare/minitab.

 

Recommended literature that can be useful to read as a supplement to the book is Box, Hunter og Hunter (1978, 2005): "Statistics for Experimenters". Wiley.

Lectures

Tuesdays 10.15 – 12.00 MA21

Fridays    08.15 – 10.00 R3

Preliminary curriculum, lecturing and progress plan

Week

Topic

Chapter in 8th Edition

(7th Edition in [.] when different)

Exercise

34

Introduction, motivation and repetition. Descriptive measures and graphs. Normal plot.

(1-10) Particularly 8.1-8.7

1

35

Two-sample case. Comparing variances. F-distribution.

Simple linear regression,

8.8, 9.8, 9.13, 10.8, 10.13, 10.18.

 

(11.1-11.5) 11.6 – 11.12

 

2-3

36-37

Multiple linear regression

12.1 -12.7, 12.9-12.11

4-5

38-40

2k experiment and fractions thereof

BHH 10 og 12
Alternatively, WMMY 15

6-8

41-42

Analysis of variance

13.1-13.4,13.6,13.8-13.10,13.13 [13.12],13.15 [13.14]
14.1-14.4

9-10

43

Statistical process control

17.1-17.5

11

44-45

Chi-square tests and Contingency tables

10.14-10.16

12

45-46

Non parametric statistics

16.1-16.3 [16.1-16.4]

12

46

Approximation of expectation and variance

 

 

47

Repetition

 

 

Exercises

Mondays 15.15 – 16.00 in MA21 or a computer lab – MAY BE CHANGED - (see information on EXERCISE WEBSITE from week to week.

The teaching assistent will on this hours be available for supervising in doing exercises or answer other relevant questions.
Exercises in order to get a deeper insight into lectured material are given each week. These can be found on the EXERCISE WEBSITE. These exercises are voluntary but are a part of the curriculum. Solutions will be put on the exercise website.
 
Most of the exercises require the use of the statistical software MINITAB, see for instance http://www.ntnu.no/itea.info/programvare/minitab.html. NTNU has bought an unlimited site liscence for Windows og Macintosh for installation of MINITAB on campus and private machines of students and faculty. MINITAB is also available on several of the data rooms on campus.

One of the exercises is compulsory and will be more comprehensive. Here the students shall plan, perform,analyse and report a self selected experiment or do another statistical analyis that need to be approved by the teacher in the course. This will count 20 % on the exam. A description and updated information on this exercise is given here (English). An older version in Norwegian is also available. Some examples of earlier experiments performed in the course are also given there. Some examples of experiments conducted in a course in USA is given here.

 

The obligatory exercise will start at the end of Week 40 (October 3), and deadline for submitting the report Is November 7 at 12 Noon. There will be no new exercises in Weeks 41 and 42 (October 6 and 13), instead you may consult the teaching assistant in the regular exercise hours these weeks. There will also be scheduled additional times in these weeks for consulting with teaching assistant and lecturer.

 

Exam

Exam: Friday December 5, 2008 at 09:00. Written exam. Durance 4 hours.
Obligatory exercise counts 20% of the exam mark.
Permitted help: All printed or hand written material. Simple, specified calculator.

 

Earlier exams in TMA4255 and similar courses:

Exam May 2007 is found in Norwegian here and in English here. Solution is  found here.
Exam May 2006 w/solution is here (in Norwegian)
Exam June 2005 w/solution is here (in Norwegian)
Exam August 2005 is here, with solution here (all in Norwegian)
Exam December 2005 is here with last pages of the solution here (all in Norwegian)
Some earlier exams w/solution are found on this old homepage (all in Norwegian).

Progress

21.10 Will sum up 13.13 and then start chapter 17, Statistical process control.

 

17.10 Went through main topics from Ch. 14 via slides. Then did 13.13 (random effects models).

A proof of Theorem 13.2 can be downloaded here.

 

14.10 Will take 13.8-13.10 (13.11-13-12 are not in curriculum). Jump to chapter 14 (will take 13.13 later). Same slides as for 07.10.

 

10.10 Continued in chapter 13: 13.4 and 13.6 (13.5 and 13.7 are not in curriculum). Same slides as for 07.10.

 

07.10 One-way ANOVA: Chapter 13.1-13.3. Slides for chapters 13 and 14 are here: (part 2, pages 1-36).

 

03.10 Concluded the section on factorial designs by looking at two earlier exam exercises. These are found on slides (pages 141-146).

 

30.09 Fractional factorial designs. You may download book chapter. As a preparation of your project you may also download a Chapter on DOE from the MINITAB manual.

 

26.09 Estimation of variances by using replicates. Blocking. New slides: (pages 118-140)

23.09 Continue from last time: Estimation of variances. Lenth’s method. Normal plot. Estimation by assuming effects are 0. New slides: (pages 104-117)

 

19.09 Started the Design of Experiments part. Factorial experiments analysed as regression models. 2^k experiments. Definitions of main effects, interaction effects. You may download either  Norwegian notes or  Copy from English book.  Slides from the lectures: (pages 87-103)

 

12.09 Finished Chapter 12, doing 12.10 and completed 12.11. Note: Chapter 12.7 is left to the lectures on Design of Experiments (see 19.09). Same slides as for 05.09.

 

09.09.Completed 12.4, then 12.6, 12.9, and Cp from 12.11.Same slides as for 05.09.

 

05.09 Prediction (Ch. 11.6). Multiple linear regression (Chapter 12): Went through 12.1-12.3 and start of 12.4, plus prediction in 12.5. Slides can be downloaded here (pages 61-86)

 

02.09 Linear regression – Chapter 11.1-11.5 Slides can be downloaded here (pages 42-60)

 

29.08 Comparison of variances for two samples. Fisher-distribution. F-test and confidence interval for sigma1/sigma2. Slides as for 26.08.

 

26.08 One-sample and two-sample t-tests. Pooled variance estimator. Slides can be downloaded here (pages 26-41)

 

22.08 Descriptive measures and graphs (e.g. normal plot). One sample from normal distribution: sampling distributions, inference with known sigma. Used same slides as last time.

 

19.08 Introduction and repetition of probability theory and some very basic concepts of statistical models and statistical inference. Slides can be downloaded here (pages 1-25)

Notes

Here are some examples that may be used in the course (some of this is unfortuantely in Norwegian). The theory presented in the note about factorial experiments and block division (the first note below), is also found in chapter 15 of the main book, or in chapter 10 of Box, Hunter & Hunter. Thereafter is a note which could replace chapter 12 in Box, Hunter & Hunter.

Notat om faktorielle forsøk og blokkdeling

Notat om fraksjonelle forsøk

Notat om rett bruk av statistikk

Notat om
Variansstabiliserande transformasjonar


Approksimasjon av forventing og varians
Eksempel i ikkje parametrisk statistikk
Kulelagereksempelet
Renseanlegget
Blodlegemer og Augefarge
Forsøksplanlegging med senterpunkt, CCD
Kontrollkort
Spenning av lagringsbatteri
Hardheit til metallgjenstand
Koparinnhald i bronslegeringar
Ph-verdi fiskevatn
Regresjon og Anscombe's datasett
Enkel lineær regresjon
Stivheit og tettleik til treprodukt
Eksempel på approksimasjon av forventing og varians
Drivstoffeksempelet
Industrieksempelet og Illustrasjon av sentralgrenseteoremt

 

Last updated: 17 October 2008 at 16:08