A method for optimizing multivariable calibrations

摘要

The present invention is a method for optimizing multivariate calibrations such as those used on FT-IR analyzers. The method involves the selection of an optimum subset of samples to use in the calibration form a larger set of available samples. The subset selection is done so as to minimize the bias of the resulting calibration while simultaneously maintaining acceptable standard errors and ensuring maximal range for the model.