Queen Mary and Westfield College
School of Mathematical Sciences
STATISTICS SEMINAR: DESIGN OF EXPERIMENTS
All are welcome
The talks are held at 16.30, all in the Mathematics Seminar Room (103)
on Level 1, Mathematics Building, Queen Mary and Westfield College.
Tea and coffee are available in the Mathematics Common Room (102)
from 15.00.
The nearest underground station is Stepney Green.
Turn left at the exit and walk 400 yards.
________________________________________________________________________
DATE SPEAKER TITLE
------------------------------------------------------------------------
21 Oct 1999 Eva Riccomagno An application of Groebner
EURANDOM bases in experimental design.
The Netherlands Algebraic Statistics.
4 Nov 1999 Paul Nelson TBA
SmithKline Beecham
18 Nov 1999 Barbara Bogacka Non-Linear Design Problem
School of Mathematical in a Chemical Kinetic Model
Sciences, QMW with Non-constant Error
Variance
2 Dec 1999 Roger Payene A Flexible Toolkit
Statistics Department for Experimental Design
IACR-Rothamsted
-----------------------------------------------------------------------
For more information ask:
Barbara Bogacka
School of Mathematical Sciences
Queen Mary and Westfield College
Mile End Road
London E1 4NS
Tel: 020 7882 5497
e-mail: [log in to unmask]
---------------------------------------------
The seminar information is kept on:
http://www.maths.qmw.ac.uk/~rab/seminars.html
______________________________________________________________________
A B S T R A C T S
______________________________________________________________________
Eva Riccomagno
An application of Groebner bases in experimental design.
Algebraic Statistics.
In the first and main part of this talk we introduce the main
idea behind the application of Groebner basis to identification of linear
polynomial models. The focus is on the attempt to algebraise and,
to automatise to some extent, the procedure to determine an
identifiable model given a design. Some examples are presented where
computational algebraic techniques are coupled with more standard
statistical analysis. Next the possibility to extend the basic theory
to more than one identifiable model and possible model selection
criteria are discussed. The notion of fan of a design is introduced
naturally. Finally a glimpse into the application in non-linear theory
is presented.
______________________________________________________________________
Barbara Bogacka
(joint work with Francis Wright)
Non-Linear Design Problem in a Chemical Kinetic Model
with Non-Constant Error Variance
We consider a design problem in non-linear regression, specifically in
a chemical kinetic model. We investigate the influence of the
dispersion structure of the random errors of observation on the design
and its efficiency. We find that there are two kinds of designs,
depending on the model parameters: a unique or a non-unique solution to
the maximisation of a design optimality criterion. We investigate a wide
range of statistical models for a simple chemical reaction and discuss
the possible solutions to the design problem. We indicate the advantage
of statistical approach to the determination of the kinetic parameters
as opposed to the standard methods used in chemistry, based on the
knowledge on the reaction rate law. We also discuss the practical
aspects of planning an optimum experiment for the estimation of the
kinetic parameters.
_______________________________________________________________________
Roger Payne
A Flexible Toolkit for Experimental Design
Statistical software can provide considerable help and guidance in the
design of experiments. It is important, however, not to impose
limitations over the range or types of experiments that are supported
by allowing choices only from a fixed set of repertoires.
In this talk I shall describe the philosophy that lies behind the Genstat
design system. This supports a range of standard generators,
such as design keys, Hadamard matrices and Alpha arrays. Files of
generators are provided to cover various standard designs, and these
are used to form interactive menus that allow users to select a design,
to name and generate its factors, print a plan and then (where appropriate)
produce a dummy analysis of variance.
Further facilities allow designs and data forms to be displayed, and simple
designs to be combined into more complicated arrangements.
Algorithms are available to form generators for new designs, and these
can be added to the design files to become an integral part of the
system. Examples will be given to illustrate the framework and
underlying methodology, and to show how the various facilities
combine to provide a very flexible toolkit for design and analysis of
experiments.
______________________________________________________________________
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|