Modeling of tapioca starch hydrolysis using neural networks

This paper considers the use of neural network models to represent the dynamic behavior of glucoamylase-starch system. The model studied in this research is multilayered feed-forward structure with time delay. In this paper we give preliminary results on the modeling adequacy of the neural network models. Data for the training were obtained from Bioprocess Laboratory, UTM. It was found that, the model was successfully developed using multilayered structure of neural network and the model exhibited satisfactory results. It is concluded that neural network technique is a feasible tool for model development of tapioca starch hydrolysis. Further modification of the structure such as increasing the time delay is also recommended for encouraging results.