Prediction of baicalin concentration within scutellaria based on terahertz spectroscopy by analyzing aligned principal component.

Abstract

Terahertz spectral analysis is made to study baicalin concentration within scutellaria by enlarging spectral feature. In detail, baicalin powder is added to scutellaria and the change in terahertz spectroscopy is intensified by multiplicative scattering correction and principal component analysis. It is found that multiplicative scattering correction contributes to aligning the first principal components, which allows the prediction of active ingredient ratio with increasing accuracy. The mechanism of feature change is verified from the perspective of spectrum overlapping of independent material. Moreover, by combining the assumption of linear superposition and the algorithm of principal component analysis, the error attached to experiment results is interpreted and attributed to difference between the claimed and actual material component. The study suggests that principal component analysis plays a role in analyzing superposition of independent terahertz spectrum, which is seldom noticed in practice.