Application of orthogonal space regression to calibration transfer without standards

Abstract

To transfer a calibration model in cases where the standardization samples are rare or unstable, a method based on orthogonal space regression (OSR) is proposed. It uses virtual standardization spectra to account for response changes between instruments or batches. A comparative study of the proposed OSR, piecewise direct standardization, finite impulse response, orthogonal signal correction, and model updating (MU) was conducted on both pharmaceutical tablet data and chlorogenic acid data. The results of these studies suggest that both the OSR and the MU are superior to the other transfer techniques in terms of root‐mean‐squared error of prediction and ratio of performance to interquartile distance. Moreover, OSR requires no identical standard samples, and it avoids re‐optimizing the transfer models. In conclusion, both the differences among spectra measured on different spectrometers and the differences between different batches can be corrected successfully using the OSR method. Copyright © 2013 John Wiley & Sons, Ltd.

Near-infrared calibration transfer via support vector machine and transfer learning

“…For recalibration using SVR and RMTL methods, selecting standardization samples through support vectors provides better results than selecting samples by the KS-algorithm. RMTL is comparable with the recently proposed orthogonal space regression (OSR), 28 when the standardization sets are both selected by the KS-algorithm. The RMSEP is 3.67 mg for RMTL while 3.64 mg for OSR.…”

“…The transfer from instrument 1# into instrument 2# is investigated. And the number of standardization samples is set to 24, because once the number of standardization samples reaches 24, a further increase in the number of standardizations has little effect on the performance of model updating methods 28. RMSE values of the primary PLS model and SVR model calibrated on instrument 1# with all calibration samples are shown in Table…”

Calibration model transfer in mid-infrared process analysis within situattenuated total reflectance immersion probes

“…In addition, global models tend to require a large number of calibration samples. 1,5 Alternatively, data pre-processing, such as multiplicative signal correction, 9 finite impulse response filtering, 10 orthogonal methods, 8,11,12 generalized least squares, 13 and wavelength selection can be employed to reduce the sensitivity of the resulting model to changes in conditions. 14 A down side of these methods can be that they tend to remove variation not present in all of the different conditions (or different instruments), and so can reduce the sensitivity of the model to the analyte of interest.…”

Investigation of Direct Model Transferability Using Miniature Near-Infrared Spectrometers

“…It would be difficult to perform calibration transfer in this way, especially when these instruments are placed in different locations. Other calibration transfer methods have been developed without using the transfer datasets from both instruments [23,24,25]. But unlike the commonly used methods, these methods have not been extensively studied and made easily available to general NIR users.…”