Neural Network Modeling Used As a Chemometric Tool for Kinetic Investigations

Abstract

General introduction on neural networks (NN), including its origin and chronological development are presented. The features of neural networks in terms of architecture, transfer function, training algorithm and the popular neural network packages are presented. Neural network modeling and other polynomial models applied to under stand on the effect of hydrogen ion on the oxidation of kinetic reaction of Ru(CN)₆ ^4- with MnO₄ ^– in perchloric acid medium. The data set is analyzed using the Trajan software. The obtained residual values with neural network (NN) modeling are far less than that in the polynomial model and also the calculated kNN values are better correlated with kobsd values.