Wavelength or Sensor Selection by Minimization of Prediction Error
Abstract
A method of variable selection based on prediction error is presented. An algorithm based on this approach is developed and used to select variables using a simulated data set. The results of this procedure are compared to those expected using signal averaging with a smaller set of variables. Keywords: Multivariate calibration, Chemometrics, Sensor selection, and Wavelength selection.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jul 01, 1988
- Accession Number
- ADA197231
Entities
People
- Bruce R. Kowalski
- Kenneth R. Beebe
Organizations
- University of Washington