Application of artificial neural networks for the simultaneous determination of a mixture of fluorescent dyes by synchronous fluorescence

A new technique is proposed for the simultaneous analysis of a mixture of the fluorescent dyes Fluorescein, Rhodamine 6G and Rhodamine B by two kinds of artificial neural networks (ANNs): back propagation neural networks (BPNNs) and radial basis function neural networks (RBFNNs). Both have been successfully applied to model the system. The synchronous fluorescence spectra (Delta lambda = 25.0 nm) of 52 standard solutions, with concentration 0-0.25 mug ml(-1), were used for this purpose. Compared with principal component regression and partial least squares, BPNNs and RBFNNs have proved to be very feasible strategies for solving the problem of the simultaneous determination of components of similar character. Moreover, because of their general validity, the new proposed approaches can also be applied to many other cases.