MULTIVARIATE CALIBRATION STANDARDIZATION

Multivariate calibration models are usually based on data from a large number of training set samples which have been collected over a long period of time. These models are meant to be used for an extended period. There are, however, a number of situations in which a multivariate calibration model may become invalid, for instance when the instrument is replaced, when drift in the instrument response occurs, when the measurement has to be taken at a different temperature, or when there is a change in the physical constitution of the samples. Multivariate calibration standardization enables one to efficiently correct for the differences between these situations and thereby eliminate the need for a full recalibration. In this paper several standardization strategies and methods, as well as some problems related to the choice of the standardization samples, are discussed.