Coding method for the study of the intrinsic mechanism of spectral analysis

Abstract

In view of many factors that influence the in situ real-time spectral measurements, various correction and modeling methods have been applied on spectral analysis. However, the intrinsic mechanism of these methods in actual applications was not usually demonstrated, so chance correlations could be unavoidable. This paper presents a new method for coding the absorbance spectrum of each component in a multi-component compound to quantize the size relation between different absorbance of various components. To describe the implementation process, spectra of the compound with three components were developed and partial least-square regression was used to construct calibration models. The results verify the feasibility of the coding method used for studying the intrinsic mechanism of spectral analysis. The new method provides a way to analyze the influence from different components of the object qualitatively and quantitatively, and it can help us to grasp the action mechanism of these correction and modeling methods. Based on this, we can choose and design a suitable method to enhance the accuracy and reliability of spectral analysis.