TY - CHAP AU - Havel, J. AU - Peña-Méndez, E.M. AU - Rojas-Hernández, A. T1 - Artificial Neural Networks in Capillary Electrophoresis LA - eng PY - 2013 SP - 77 EP - 93 T2 - Capillary Electrophoresis and Microchip Capillary Electrophoresis: Principles, Applications, and Limitations SN - 9780470572177 PB - John Wiley and Sons AB - Several methods and approaches have been developed and are in use to optimize capillary electrophoresis, micellar electrokinetic chromatography, and/or electro chromatography. The search for parameters for which the optimal or "the best" separation is achieved is sometimes straightforward and simple, but often requires exploration of a high number of parameters. Classical "single variable," or relaxation method when just one parameter is changed at a time, should be omitted as the convergence is slow or the process does not even converge at all. Multivariate chemometrics techniques should be used. Applying suitable experimental design (ED) such as factorial design or central composite design is recommended as the number of experiments can be substantially reduced. There might be an advantage modeling the data obtained using methods such as artificial neural networks (ANNs) even when separation principles are not clear or physicochemical parameters are not available. Basic principles of ED, ANN, and a combination of both (ED-ANN) approaches are given herein, as well as critical overviews of other ANN applications. © 2013 John Wiley & Sons, Inc. DO - 10.1002/9781118530009.ch5 UR - https://portalciencia.ull.es/documentos/5e3c38cd29995246bbf5eed9 DP - Dialnet - Portal de la Investigación ER -