Dear all!
I would like to remind you that TOMMOROW the JOINT ST ANDREWS/HIGHLAND
RSS GROUP MEETING takes place. Note that the sequence of our two
speakers has changed compared to the first announcement. You can still
register for the dinner after the meeting until 5pm TODAY, Tuesday
2/05/2006 (e-mail [log in to unmask])!
SPEAKERS: Frank Critchley (Open University), Simon Wood (Bath)
DATE: WEDNESDAY 03/05/2006
TIME: 3pm
VENUE: Lecture Theatre D, Mathematical Institute, North Haugh, St Andrews
(for updates please go to:
http://www.mcs.st-and.ac.uk/StatsSeminars/index.shtml,
a map of the university can be found at:
http://www.st-andrews.ac.uk/map/map.pdf)
SCHEDULE (abstracts given below):
3.00pm Simon Wood (University of Bath):
"Calanus in the North Atlantic: a simple approach to fitting a complex
model."
4.00pm: Tea (in the Staff Room)
4.30: Frank Critchley (Open University): "Principal Axis Analysis "
The meeting will be followed at 6.30 p.m. by a 2-course meal at the
Doll's House Restaurant in Church Square, St Andrews. Those wishing to
come to the meal, please inform Peter Jupp, preferably by email
([log in to unmask]) before the end of Monday 1 May (updates on the
arrangement for the meal can also be found at:
http://www.mcs.st-and.ac.uk/StatsSeminars/index.shtml)
--------------------------------------------------------------------------------------------------------------------------------------------------------------
Calanus in the North Atlantic: a simple approach to fitting a complex
model (Simon Wood, Department of Mathematical Sciences, University of
Bath, Joint work with W. Horbelt, D.C. Speirs, M.R. Heath and W.S.C. Gurney)
Calanus finmarchicus is the dominant zooplankton species over much of
the North Atlantic. It has an unusual life-cycle involving overwintering
in very deep waters at very cold temperatures and then re-ascending to
breed in the spring. Because of the considerable distances that
individual calanus are transported during their lives, calanus
population dynamics can only be properly understood at the scale of the
whole North Atlantic. This involves the formulation of physically driven
spatially explicit population dynamic models describing the whole north
atlantic population. However, given the complexity of the calanus
life-cycle, key model parameters can only be estimated by treating this
complex model as a statistical model and fitting it to data on calanus
abundance. This talk describes an approach to doing this, involving a
(parallelized) hybrid of finite differencing and autodifferentiation to
obtain derivatives of the model predictions w.r.t. the parameters,
coupled with a modified Gauss-Newton fitting method. Despite the model's
complexity, the approach allows a model fitting, checking and refinement
cycle rather similar to that used in linear modelling, for example.
-----------------------------------------------------------------------
Principal Axis Analysis
(Frank Critchley(1), Ana Pires(2) and Conceição Amado(2); Open
University, UK(1) and IST, Lisbon(2))
Principal axis analysis rotates standardised principal components to
optimally detect subgroup structure, rotation being based on
preferred directions in the spherised data. As such, it is a
computationally efficient method of exploratory data analysis,
particularly well-suited to detecting mixtures of elliptically
contoured distributions. The ability of principal components itself
to perform as a cluster analysis method on some occasions, but not
others, is explained and illustrated. Links with a number of related
multivariate methods are explored. Examples are given throughout.
Further developments are briefly indicated. Overall, principal axis
analysis exemplifies the maxim: 'two decompositions are better than
one'.
--
***********************************************************************************
Dr Claus-D. Mayer | http://www.bioss.ac.uk
Biomathematics & Statistics Scotland | email: [log in to unmask]
Rowett Research Institute | Telephone: +44 (0) 1224 716652
Aberdeen AB21 9SB, Scotland, UK. | Fax: +44 (0) 1224 715349
***********************************************************************************
|