RANDOM CORRELATION IN VARIABLE SELECTION FOR MULTIVARIATE CALIBRATIONWITH A GENETIC ALGORITHM

Citation
D. Jouanrimbaud et al., RANDOM CORRELATION IN VARIABLE SELECTION FOR MULTIVARIATE CALIBRATIONWITH A GENETIC ALGORITHM, Chemometrics and intelligent laboratory systems, 35(2), 1996, pp. 213-220
Citations number
15
Categorie Soggetti
Computer Application, Chemistry & Engineering","Instument & Instrumentation","Chemistry Analytical","Computer Science Artificial Intelligence","Robotics & Automatic Control
ISSN journal
01697439
Volume
35
Issue
2
Year of publication
1996
Pages
213 - 220
Database
ISI
SICI code
0169-7439(1996)35:2<213:RCIVSF>2.0.ZU;2-U
Abstract
The importance of the validation step in multiple linear regression of near-infrared spectroscopic data, after selection of wavelengths by a genetic algorithm, is investigated with the use of random variables. It is shown that in spite of a careful validation procedure, the GA ca n still select irrelevant variables. The effect is greatly reduced by applying a forward selection in the subsets selected by the genetic al gorithm.