An equivalence relation between parallel calibration and principal component regression

Multivariate calibration and prediction, when there are few observations and many variables, is an important and unresolved problem in chemometrics. Recently a technique called paralle calibration was introduced as a new and easy to understand calibation method. In this paper we show that parallel calibration is equivalent to a form of principal component regression and to generalized inverse regression. Copyright (C) 2002 John Wiley Sons, Ltd.