Correction of temperature-induced spectral variations by loading space standardization.

With a view to maintaining the validity of multivariate calibration models for chemical processes affected by temperature fluctuations, loading space standardization (LSS) is proposed. Through the application of LSS, multivariate calibration models built at temperatures other than those of the test samples can provide predictions with an accuracy comparable to the results obtained at a constant temperature. Compared with other methods, designed for the same purpose, such as continuous piecewise direct standardization, LSS has the advantages of straightforward implementation and good performance. The methodology was applied to shortwave NIR spectral data sets measured at different temperatures. The results showed that LSS can effectively remove the influence of temperature variations on the spectra and maintain the predictive abilities of the multivariate calibration models.