Evaluating Dye Concentration in Bi-Component Solution by PCA-MPR and PCA-ANN Techniques

This paper studies the application of principal component analysis, multiple polynomial regression, and artificial neural network techniques to the quantitative analysis of binary mixture of dye solution. The binary mixtures of three textile dyes including blue, red and yellow hues were analyzed by PCA-Multiple polynomial regression and PCA-artificial neural network methods. The obtained results indicate that the accuracy of PCA-artificial neural network technique is higher than PCA-Multiple polynomial regression and normal spectroscopy methods. The PCA-artificial neural network technique is applicable for dye concentration bicomponent solution with both overlapping and non-overlapping spectra. The developed method can be a practical solution to quantitative analysis of binary mixture of dye solutions with overlapping. Prog. Color Colorants Coat. 6(2013), 129-139. © Institute for Color Science and Technology.