 |
Royal Statistical Society
Statistical Computing Section
Meetings
|
High Dimensional Data Meeting
High-dimensional data arise when there are more variables (or
parameters) than observations. This meeting will provide an overview of
statistical problems and solutions involved in high-dimensional data
analysis, together with a discussion of computational issues.
Applications involving analysis of spectroscopic and micro-array data
will be presented.
"Statistical challenges in micro-array data"
Hans van Houwelingen (Department Medical Statistics, Leiden University Medical Centre
)
Synopsis: The statistical challenges encountered in the analysis of micro-array
data depend on the purpose of the data collection. That can either be to
discover pathways by clustering genes, to find out which genes are
affected by some intervention or to predict some clinical outcome. A
review is given of challenges and possible solutions.
- The slides for this talk are available in pdf format.
"Quantitative and qualitative calibration using high-dimensional
spectroscopic data"
Tom Fearn (Department of Statistical Science, University College London
)
Synopsis: The advent of micro-arrays has awakened statisticians to the existence
of `small n, large p' problems. However, such problems have been tackled
in spectroscopy for many years, providing the biggest success story for
`chemometrics'. The talk will describe some spectroscopic examples.
- The slides for this talk are available in pdf format.
"Efficient Bayesian approaches for addressing model uncertainty in the
face of many potential predictors"
Chris Holmes (Department of Mathematics, Imperial College
)
Synopsis: We discuss Bayesian approaches to dimensionality reduction in
generalised regression settings, including GLMs and nonlinear regression
spline models. An emphasis is placed on efficient computational
procedures that can provide solutions in realistic time. The methods,
including variable selection and orthogonal projection techniques, are
illustrated on data sets in genomics.
Date & Time
Wednesday 5th February 2003 at 2:00 - 5:15 pm
Place
Errol Street
Refreshments
There will be tea and biscuits served at 3:45 pm.
Enquiries
Phil Bowtell (01784 498374)
bowtell.p@pg.com
The Royal Statistical Society
Statistical Computing Section home page