Course
at a Glance/Objectives
The course aims at presenting the statistical theory of design of
experiments. The basic question is at which
settings to observe a system in order to estimate, or at least
identify, some specific model parameters.
Emphasis will be given to the study of optimal designs, where
optimality refers to specific statistical
properties of the estimates. If time allows, a recently developed
algebraic encoding of designs supported on
polynomials will be introduced.
There will be eight meetings of two and half hour each roughly.
First
meeting: 7th October at 4pm at Department of Mathematics (room to be
announced)
1. General concepts, controlling for bias and variation, interplay
between design and analysis.
2. Factorial designs and aliasing.
3. Basics on computational commutativa algebra, polynomial
representations of factorial designs.
4. Optimal designs for linear regression models.
5. Optimality criteria.
6. Algorithms for design construction.
7. Space filling designs.
8. Adaptive and sequential designs.
Organization
The course develops in about 20 hours will be in English.
Examination
Reference
textbook
D.R. Cox and N. Reid (2000), The theory of the design of experiment,
Chapman & Hall/CRC.
Design and Analysis of Experiments (Springer Texts in Statistics) by
Angela M. Dean and Daniel Voss (Dec 21, 2000)
Optimal Design of Experiments: A Case Study Approach by Peter Goos and
Bradley Jones (Aug 15, 2011)
Statistics for Experimenters: Design, Innovation, and Discovery , 2nd
Edition by George E. P. Box, J. Stuart Hunter and William G. Hunter
(May 31, 2005)
7 October 16-18.30 |
Introduction to design and basic
principles and examples and general linear models.
|
Eva |
|
17 October 15-17.30 | Examples and Gauss Markov theorem Dean Voss Chapter 2 Wu Hamada Chapter 1.4 |
Sara | |
21 October 16-18.30 | Factorial
designs, main factors and interactions and their estimation, mainly for
binary designs, orthogonal X. Ideas on orthogonal designs for more than
two level factors. Blocking factorial designs. CoxReid Chapter 5.1-5.5 Notes for Atesh |
Atesh | |
28 October 16-18.30 |
More on fractional factorial and
other design systems CoxReid Chapter 5.6 fractional factorial, 6.2.1 simple confouding, 6.2.2 double confounding, 6.3.4 supersaturated systems, 6.4.1 split plot designs, 6.6.4 mixture designs, 7.7 Latin hypercube as space filling designs. Read 6.8! Please check the topics also in the two other volumes. |
Mahad
|
|
13 November 14-18.30 |
Topic 1: Response surface models and designs Main reference: Wu Hamada Chapter 9 (excluded 9.5 and 9.6) Dean Voss Chapter 8 Cox Reid 6.6 Topic 2: Optimal design theory There are two sets of notes on dropbox Cox Reid Chapter 7 2.1 Introduction and optimality criteria, continuous designs and information matrix and its properties (Eva) very old notes Chapters 1-2-3, Cox Reid Paragraphs 7.1, 7.2, 7.3, 7.4 2.2. General equivalence theorem and algorithms (Loris) very old notes Chapters 4-5-6, Cox Reid Paragraphs 7.3, 7.5 Topic 3: Anova There are two sets of notes on dropbox. Haneef also is looking for his own material. |
Gianluca Eva Loris Haneef |
|
26 November 15-17.30 |
Introduction to Robust parameter design Main reference: Wu Hamada Chapter 10 |
Abdul Talha |
|
9 December 14- 17 |
Case studies in optimal design: Chapter 1 in Goos Jones - Eva Chapter 2 in Goos Jones - Mariacarla Chapter 4 in Goos Jones - Maria Laura A proposal for a design of experiment |
Mariacarla Maria Laura Amirreza |