A comparison of neural networks and partial least squares for deconvoluting fluorescence spectra

Abstract

This article compares backpropagation neural networks (BNN) with partial least squares (PLS) techniques in terms of their ability to deconvolute fluorescence spectra. Both actual experimental and simulated spectral data are studied for 2 binary systems. These systems consist of mixtures of tryptophan and tyrosine, and NADH and tryptophan over a total concentration range of 10(-7) to 10(-4) M. It is shown that BNN is superior to PLS for both systems.

Neural networks for kinetic parameters determination, signal filtering and deconvolution in thermal analysis

“…er(x) = I(T-o)/r I (6) where T represents the target value of the parameter x and O the output estimated by the network. As an example, the relative errors can be found in Table 4 for data with and without noise.…”

“…Some workers have already reported on several applications in which neural networks have been used with success for deconvolution purposes in various fields of research [6][7][8][9], but not for calorimetric field. Furthermore, in addition to the high capability of neural networks to deal with nonlinear problems, it is known that they are highly tolerant to noise, and numerous authors have reported on very good robustness in the presence of noisy signals [10][11][12].…”

Determination of inhibition in the enzymatic hydrolysis of cellobiose using hybrid neural modeling

“…Specifically, in relation to biotechnological processes, several studies can be found in the literature, such as the description of the α-amilase inactivation (Geeraerd et al, 1998), the prediction of the final concentration of ethanol in a batch fermentation process (Saucedo et at., 1994) and as a soft-sensor (McAvoy et al, 1992).…”

Computational neural networks for mapping calorimetric data: Application of feed-forward neural networks to kinetic parameters determination and signals filtering

“…These effects are then recorded by the apparatus as for an experimental study and will provide a way of evaluating the transfer function, for a transformation occurring during a certain interval of time depending on its kinetic parameters, and not only applicable to a Dirac pulse or a unit step function. Some workers have already reported on several applications in which neural networks have been used with success for deconvolution purposes in various fields of research [9][10][11][12], but not for calorimetric field. Furthermore, in addition to the high capability of neural networks to deal with nonlinear problems, it is known that they are highly tolerant to noise, and numerous authors have reported on very good robustness in the presence of noisy signals [13][14][15].…”

“…This occurs expecially when multiple linear regression methods for kinetic models with more than one exponent are used. Nevertheless, many comparisons have been made between artificial neural networks and conventional multivariate statistical methods [9,[16][17][18][19][20], such as multiple linear regression methods, principal components regression or partial least square regression methods, and superior results were generally obtained for neural networks.…”