Application of multivariate calibration and artificial neural networks to simultaneous kinetic-spectrophotometric determination of carbamate pesticides

Neural Computation Methods in the Determination of Fungicides

“…Applications in cited papers include those of Suárez Araujo and García Báez (Suárez Araujo, 2006) (García Báez, 2010) in which they identify fungicides in mixtures of up to three and four different fungicides using ANNs and neural network ensembles. Research by Guiberteau and collaborators have solved ternary mixtures of pesticides with similar molecular structure (Guiberteau, 2001), Baoxin Li (Li, 2006) identify three organophosphorated pesticides, Istamboulie solves insecticides (Georges, 2009), and Wan Chuanhao (Wan, 2000), who proposes the use of neural computation to identify pesticides from the family of carbamates, or the developments by Yongnian Ni and collaborators (Ni, 2004), that describe the RBF-ANN method and the PC-RBF-ANN method as the best at detecting mixtures of three pesticides of this same family. On the other hand, work by Karl-Heinz Ott (Karl-Heinz, 2003) shows an example of neural networks that determine the action mode for a large number of herbicides.…”

A combined artificial neural network/residual bilinearization approach for obtaining the second‐order advantage from three‐way non‐linear data

“…y 2 ðtÞ ¼ y R r 1 ½1 À expðÀr 2 tÞ 1 þ r 1 (13) where y R is the initial concentration of R (taken as 0.01), y 1 (t) and y 2 (t) are the concentrations of P A and P B , respectively, at For the simulations, the concentrations of P A and P B were calculated at 19 different times, in the range 1-19, using Equations (12) and (13). Two column vectors y 1i and y 2i (size 19  1), were constructed for the ith calibration sample, and converted to (g 1i c 1 ) and (g 2i c 2 ), respectively, where c 1 and c 2 are unit-length normalized time profiles, and g 1i and g 1i are scaling factors.…”

“…Only a few applications of neural networks to non-linear spectroscopic second-order data are known: appropriate examples are the kinetic-spectrophotometric determination of three carbamate pesticides [13], the correlation between two-dimensional nuclear magnetic resonance data with the composition and properties of oil samples [14], and the monitoring of fermentation processes [15]. Little is known, however, as to whether the second-order advantage can be obtained from higher order information in the presence of unsuspected sample components, in any of the two modes depicted in Figure 1, or in additional, yet unimagined ways.…”

Reduced‐order model for monitoring spectroscopic and chromatographic polymer properties

“…The goal of these studies has been to elucidate the dynamics governing the behavior of a process. For example, a combination of mass action kinetic model and artificial neural network has recently been applied to monitor different levels of pesticide contents in fruit and vegetable samples, whose levels were obtained by using spectrophotometry techniques [1]. Similar studies focused on (a) identifying and estimating a unique set of rate coefficients by simultaneously fitting a set of measurements [2][3][4], (b) understanding how changes in measurements affect model parameter values and the initial rate of reactions [5], (c) decreasing parameter search space by including parametric non-negativity constraints in the parameter estimation [6] and (d) understanding the effect of the initial guesses on an unconstrained optimization approach to parameter [4] (solid lines = model prediction; triangles = measurements).…”