Estimation of component spectral curves from unknown mixture spectra
- 15 June 1984
- journal article
- Published by Optica Publishing Group in Applied Optics
- Vol. 23 (12), 1955-1959
- https://doi.org/10.1364/ao.23.001955
Abstract
In this paper, we propose a method which uniquely determines a set of single curves, each as an estimate of a component spectrum. No reference spectrum from a library is necessary; the spectrum set of mixtures of unknown components with various concentrations is used for the estimation of component spectral curves. The method is based on entropy minimization. In comparison with an earlier method [Appl. Opt. 22, 3599 (1983)], which gives the bands of the possible component spectra, this method has the advantage of providing a unique estimation of the component spectra, which helps chemists with quantitative analysis and further mathematical processing. Two experimental results demonstrate the effectiveness of the method: for infrared absorption spectra of xylene isomers and visible absorption spectra of dyes.Keywords
This publication has 8 references indexed in Scilit:
- Constrained nonlinear method for estimating component spectra from multicomponent mixturesApplied Optics, 1983
- Computer-Assisted Spectral Identification of Unknown MixturesApplied Spectroscopy, 1982
- Noise Reduction via Factor Analysis in FT-IR SpectraApplied Spectroscopy, 1982
- Determination of the number of factors and the experimental error in a data matrixAnalytical Chemistry, 1977
- Theory of error in factor analysisAnalytical Chemistry, 1977
- Self Modeling Curve ResolutionTechnometrics, 1971
- Computer Program for Multicomponent Spectrum Analysis Using Least Squares Method.Analytical Chemistry, 1965
- Smoothing and Differentiation of Data by Simplified Least Squares Procedures.Analytical Chemistry, 1964