EPSRC Reference: |
GR/K73343/01 |
Title: |
FEATURE SELECTION WITH MANY VARIABLES |
Principal Investigator: |
Brown, Professor P |
Other Investigators: |
|
Researcher Co-Investigators: |
|
Project Partners: |
|
Department: |
Sch of Maths Statistics & Actuarial Sci |
Organisation: |
University of Kent |
Scheme: |
Standard Research (Pre-FEC) |
Starts: |
01 March 1996 |
Ends: |
31 August 1999 |
Value (£): |
115,916
|
EPSRC Research Topic Classifications: |
Statistics & Appl. Probability |
|
|
EPSRC Industrial Sector Classifications: |
Pharmaceuticals and Biotechnology |
|
|
Related Grants: |
|
Panel History: |
|
Summary on Grant Application Form |
Modern electronic instrumentation allows the rapid availability of high dimensional data structures. The number of data points is often relatively few. Thus in near-infrared (NIR) spectroscopy there may be a few hundred samples with each sample consisting of a spectrum of absorbances at perhaps 1000 wavelengths. These are to be used to ascertain the different chemical compositions (calibration) or groupings (discrimination) of the samples. In order to create generalisable models from such training data we investigate multivariate variable selection as a means of data reduction. The methods investigated will incorporate prior knowledge in a generic way so as to form robust selections. Information or logarithmic probability utility will provide one root. Specialised applications are to wavelength selection on the one hand and selection of derived variables as in orthogonal series decomposition (principal components, splines, Fourier, wavelets) or in neural networks on the other hand. We propose to study a range of approaches both Bayesian and non-Bayesian, for linear and non-linear modelling, with particular reference to chemometric problems. Pfizer Ltd will provide data and advice on blending and reaction monitoring.
|
Key Findings |
This information can now be found on Gateway to Research (GtR) http://gtr.rcuk.ac.uk
|
Potential use in non-academic contexts |
This information can now be found on Gateway to Research (GtR) http://gtr.rcuk.ac.uk
|
Impacts |
Description |
This information can now be found on Gateway to Research (GtR) http://gtr.rcuk.ac.uk |
Summary |
|
Date Materialised |
|
|
Sectors submitted by the Researcher |
This information can now be found on Gateway to Research (GtR) http://gtr.rcuk.ac.uk
|
Project URL: |
|
Further Information: |
|
Organisation Website: |
http://www.kent.ac.uk |