ABSTRACT
This chapter describes calibration techniques related to univariate and multiregression methods, focusing on principal component regression, partial least squares (PLS) regression, and presents a general overview of other multivariate techniques. Calibration involves using the data gathered from the system by means of diagnostic techniques and using them to predict one or more of its property or underlying parameter. The application of linear modelling is not limited to spectroscopy and can be employed in a wide range of applications. One of the methods to compress the information included in the data-set and to reduce the correlation among variables is to calculate the principal component of the data-set before applying the regression. The development of PLS started due to the different methodologies that one can use to select the components to be employed in the PCR. The idea was that people can employ factors in common by their x and y.
