A solution to spurious peaks in near-infrared spectroscopy model transfer

Abstract

The invention discloses a method for solving spurious peaks in near-infrared spectrum model transfer. In the method, firstly, the absorbance data of each wave number point of the source spectrometer is selected from the window near the corresponding wave number point of the target spectrometer; and then the spectral data in the window of the target spectrometer Perform principal component analysis to obtain its principal component components, and establish a linear model from it to the absorbance data of the source spectrometer; again, use the linear regression solution method with coefficient norm penalty for the data of each wave number point of the source spectrometer to establish the full spectrum Spectral shift model; finally, by adjusting the penalty parameter, the degree of peak removal of the shifted spectrum can be controlled. The invention uses a linear regression method with a coefficient norm penalty to establish a spectrum transfer model, and compared with a traditional transfer model based on a partial least square method, the spectrum is smoother and more robust.

Description

technical field [0001] The invention belongs to the field of rapid analysis of properties based on near-infrared spectroscopy, and specifically relates to a method for solving spurious peaks in near-infrared spectroscopy model transfer. Background technique [0002] In the field of rapid property analysis based on near-infrared spectroscopy, the problem of model transfer is often involved. Using a calibration model established by a spectrometer, even if it is another instrument of the same model, directly applying the calibration model may produce large prediction deviations. Usually, this problem is solved by establishing a spectral correspondence between the target spectrometer and the source spectrometer, which is called model transfer. [0003] Currently, this transfer model is usually established as a partial least squares (PLS) model from the target spectrometer data window to the source spectrometer data points. However, in practical applications, the PLS model is p...

AI Technical Summary

Problems solved by technology

[0004] Aiming at the problem that spooky peaks are prone to appear in the current PLS-based spectral model transfer, the present invention proposes a solution to the spooky peaks in the near-infrared spectral model transfer. The main steps include:

Images