Using RBF-based artificial neural network to optimization of betulinic acid ester enzymatic synthesis

Abstract

Modeling of reaction conditions in the enzymatic synthesis of betulinic acid ester with multi-layer perceptron (MLP) which is a class of feed forward artificial neural network (ANN) has been done. Error to test data depends on not only trained method but also number of neuron in hidden layer. Often, it is much more of error to training data. In this manuscript Radial Basis Function (RBF) has been utilized. It determines the number of neuron due to desired error. In appropriate ANN, error for training data and test data are close. Therefore RBF’s parameters are set by multi objective meta-heuristic algorithm, in a way that, mentioned purposes have been obtained. The results of this method will be compared with previous methods.