FULLY FUNDED CASE PHD STUDENTSHIP AT THE UNIVERSITY OF BRISTOL,
IN CONJUNCTION WITH SHELL RESEARCH

Local autocovariance methods for time series analysis,
with change-point applications

Applications are invited for a three year CASE studentship within the
School of Mathematics at the University of Bristol, jointly funded by the
EPSRC and Shell Research Ltd.  The studentship will focus on the development
of new tools for the analysis of non-stationary time series with
Professor Guy
Nason as adviser. The anticipated start date is October 2003.

FURTHER DETAILS

http://www.stats.bris.ac.uk/~guy/Jobs/Shellcase.htm

Further details about the area of study and what is on offer are
provided below.
Additional details may obtained by contacting:

Professor Guy Nason
Professor of Statistics
Department of Mathematics
University of Bristol
University Walk
BRISTOL BS8 1TW

E-mail:        [log in to unmask]
Web-site:      http://www.stats.bris.ac.uk/~guy
Telephone:     0117 9288633

If you would like to apply for the studentship, please send a brief
letter of
application including your CV to Prof. Nason at the above address.

The closing date for applications is June 27, 2003.

FINANCIAL DETAILS

If you are a UK student.

The studentship fully pays your fees and also pays a full grant.
For 2003/4 the total amount is #12000 (This is tax free and consists of
#9000 from the EPSRC and #3000 from Shell Research Ltd. The figure of #9000
is due to rise significantly. In the recent Government white paper a figure
of #12000 in 2005-6 is quoted).

If you are an EU student.

The studentship will pay your fees, but won't unfortunately pay your
living expenses (maintenance grant).

If you are an overseas student.

Sadly, the studentship won't pay anything. However, we do have other
routes to
support outstanding graduate students, please see our website
http://www.stats.bris.ac.uk/Postgrad

FURTHER DETAILS.

About the research

Large quantities of data accumulating in the retail and industrial sector
offer significant opportunities for novel developments in statistics.
This project will develop statistical methods for the modelling and analysis
of (possibly multivariate) non-stationary time series, in particular
localized
autocovariances.

There are countless opportunities for adoption of non-stationary time series
methods for change-point analysis within an organisation like Shell, with a
broad base of activities including notably large-scale manufacturing,
retail and financial services. Within the manufacturing environment,
for example, the capability to model spatio-temporal systems is of
significant
value: from modelling of industrial furnaces and heat exchangers
(relatively straightforward) to understanding highly complex production
facilities such as the catalytic cracker.

In the economic arena, detection of changes in the structure of data for
financial systems is of significant value.

About the department

The student will join a dynamic and friendly group at the University of
Bristol working in the general area of multiscale methods in statistics.
The Bristol Statistics Group is an attractive and supportive environment for
graduate students with the additional benefit of having consistently
demonstrated research excellence within statistics, covering a wide
variety of
areas in probability and statistics with specific expertise in computational
statistics.

The group benefits from a large number of graduate students, which,
together with several postdocs makes for a creative atmosphere within
which to
develop new ideas.

The Statistics Group has access to world-class University facilities as well
as excellent Departmental computational resources (for example, incoming
graduate students typically get their own high-end PC.The Department
also runs
a 20 processor Linux Farm and a 162 processor Beowulf).

The Department of Mathematics is situated in the main University precinct in
the beautiful Royal Fort Gardens. The precinct is a few minutes' walk away
from Park Street/The triangle/Queen's Road/Whiteladies road with an
interesting
mix of shops, restuarants, pubs and parks. The Department is about 15
minutes
walk away from Broadmead: a major Bristol shopping centre which
surrounds John
Wesley's chapel and also the same distance to the Harbourside with the
acclaimed @Bristol, museums and galleries. See http://www.bris.ac.uk for
more
on the University and its surrounds.

See http://www.bris.ac.uk for more on the University and its surrounds.
See the Bristol Tourist Information website
http://www.visitbristol.co.uk for
more information about the city.

About the industrial partner

Dr Idris Eckley of Shell Global Solutions (UK) will be the industrial
adviser.
Dr Eckley is a visiting fellow in the Statistics Group at Bristol
University.
Shell Global Solutions (UK) is a part of Shell Research and is based
in Chester in the north west of England. For more general information
see http://www.shellglobalsolutions.com/index.htm


Your profile

We are looking for students who have recently graduated or about to graduate
with a first class degree in Mathematics or Mathematics and Statistics. Some
previous experience of programming (e.g. FORTRAN, C, Splus, R or MATLAB) is
desirable but not essential. It is anticipated that some of the developed
scientific ideas will be implemented in C or MATLAB. Full training will
begiven if necessary, thus enhancing the student's transferrable skills.
You are motivated to develop, in time, cutting-edge research which has
applications founded on real-world problems.

What's on offer

Close interaction between yourself, the academic and industrial partner
should
provide an exciting opportunity for the student to develop novel
research which
will be applied to real industrial problems.

The student will also spend periods of time at Shell Research in Chester.
For a CASE award the total amount of time is a minimum of three months
over three years. The exact timings and periods to be determined by the
student, adviser and Shell.

The student will emerge from the studentship equipped with leading expertise
in non-stationary time series analysis, other statistical methods and tools,
experience of working for a large multinational in a research environment.
Judging by past experience this will make the student highly employable
by industry or academia.