Neural networks as a tool for modelling the retention behaviour of dihydropyridines in micellar liquid chromatography

Optimization strategy for liquid chromatography–electrospray ionization mass spectrometry methods

“…In LC, empirical modelling has been used to model retention behaviour, [4][5][6][7][8] but to our knowledge only one attempt 9 has been made to incorporate chemometric modelling in the optimization of an LC-MS method. That study involved an LC-APCI (atmospheric pressure chemical ionization)-MS interface with optimization of signal intensity rather than S/N ratio and without consideration of the chromatographic quality.…”

Micellar liquid chromatography

“…Potential advantages of the application of artificial neuronal computational networks instead of classical statistical methods for simulating the dependence of the retention time on the concentrations of surfactant and organic modifier have been demonstrated using a mixture of 27 dihydropyridines 68 or a mixture of 23 benzene derivatives and PAH. 69 The effects of the network architecture (the binding mode of individual neurons), the nodal activation function, the number of layers and the mode of the data processing were studied.…”

“…69 In the case of dihydropyridines, 1 ± 3 neurons and recurrent networks were used. 68 Different transmission functions (two sigmoidal, one logarithmic and one linear) were activated in the node of the hidden layer; of these the logarithmic function proved to be the optimum. The optimum number of nodes in the hidden layer was equal to three.…”

A QSPR Study on the GC Retention Times of a Series of Fatty, Dicarboxylic and Amino Acids by MLR and ANN

“…6,7 There are several reports on the use of neural networks in the modeling of retention behavior and optimization of conditions in micellar liquid chromatography. 8,9 Artificial neural networks have also been applied to a wide variety of chemical problems such as different QSAR studies, 10,11 prediction of 13 C NMR chemical shift, 12 selectivity coefficients of ion-selective electrodes, 13 simulation of mass spectra and modeling of ion-interaction chromatography. 14,15 (3) Partial least squares (PLS), [16][17][18][19] which is based on factor analysis, is applied where there are many variables but not enough samples or observations.…”