This paper aims at serving two purposes: firstly, it gives a quick summary
of aspects and properties of wavelets and wavelet transforms which are need
ed in order to understand how to (pre-)process data from spectrometry with
wavelet methods. Secondly, it shows on a typical example (wheat NIR spectra
l how wavelet transforms can be used in order to extract quantitative infor
mation. In contrast to other approaches in the literature, we use special t
ypes of wavelets which allow analysing finitely extended signals without in
troducing artifacts near the boundaries, and we introduce a new way of wave
let coefficient regression in order to build our chemometrical models. (C)
2000 Elsevier Science B.V. All rights reserved.