A comparison of multivariate calibration techniques applied to experimental NIR data sets Part II. Predictive ability under extrapolation conditions

The present study compares the performance of different multivariate calibr ation techniques when new samples to be predicted can fall outside the cali bration domain. Results of the calibration methods are investigated for ext rapolation of different types and various levels. The calibration methods a re applied to five near-IR data sets including difficulties often met in pr actical cases (nonlinearity, nonhomogeneity and presence of irrelevant vari ables in the set of predictors). The comparison leads to general recommenda tions about what method to use when samples requiring extrapolation can be expected in a calibration application. (C) 2001 Elsevier Science B.V. All r ights reserved.