In this paper, a novel technique for correction of background drifts is rep
orted. The method is based on the use of artificial pseudo principal compon
ents (pPC). These pPC are used for modeling the drifts of the background of
each measurement spectrum individually. They are handled like the known pr
incipal components (PC) that build up the absorption features. Simultaneous
evaluation of drifts and absorptions improve the correction ability. This
is demonstrated by means of simulated spectra as well as by experimental UV
transmissions spectra. (C) 2000 Elsevier Science B.V. All rights reserved.