透射光谱线性空间核学习建模求解多组分浓度

本文对求解多组分体系浓度的分光光度计同时测定法提供了一种建模方法.通过对多组分体系的透射光谱建立光谱线性空间,证明了多组分体系的分子势函数矩阵V_M可以转化为对角矩阵,对角矩阵的对角元是单组分势函数的块矩阵J_i.多组分体系透射光谱空间是单组分分子透射光谱函数为基函数、分子数量占比为坐标的线性组合.利用多核学习法确定各单组分体系透射波的占比权重系数,提出了用单组分浓度光量子隧穿软测量模型测定多组分浓度的测量方法.实例验证表明此方法稳定可靠,且能降低计算的复杂性.

Kernel learning modeling for solving multi-component concentrations with linear space of transmission spectra

A modeling method is proposed in this paper for simultaneous spectrophotometric determination of multicomponentsystem concentrations. A linear space of transmission spectra of the multi-component system was established,which proved that the potential function matrix VM of the multi-component system can be transformed into a diagonalmatrix, in which the diagonal elements are blocked matrix Ji of the single-component potential functions. The multicomponentsystem transmission spectral space is a linear combination of the transmission spectral functions of each singlecomponentmolecule as the basis function and the proportion of molecular numbers as the coordinate. The weight coefficientsof the transmission waves of each single-component system were determined using the multi-kernel learning method,and a measurement method for measuring multi-component concentration using a single-component concentration quantumtunneling soft measurement model was proposed. Example validation shows that this method is stable and reliable andcan reduce computational complexity.