Prediction of Probiotic Starter Concentration in Functional Food Technology Using Artificial Neural Networks

To predict probiotic starter concentration (outputs) based on specified variables inputs (paring Meal extract concentration, yeast extract concentration and glucose concentration) during a fermentation process by using artificial neural network. Samples as well as ranges of data patterns (input output) collected from Pediococcus acidilactici fermentation process to develop the artificial neural network model for a Pediococcus acidilactici concentration. Such a neural network identification process, in turn, needs some optimization methods to find the best network architecture. The parameters of interest in this multi-input, single output system that affect the Pediococcus acidilactici concentration are paring meal extracts concentration, yeast extract concentration and glucose concentration. Results suggest that artificial neural network provide an effective means of efficiently recognizing the patterns in data and accurately predicting Pediococcus acidilactici concentration based on investigating inputs, and also can be used to estimate probiotic starter concentration (outputs) based on specified variables inputs (paring Meal extract concentration, yeast extract concentration and glucose concentration) on fermentation process. Artificial neural networks effectively recognize pattern in data. Also this study develops accurate and time saving model to predict concentration of Pediococcus acidilactici. Such equations allow food industries and researchers to estimate probiotic starter concentration in relation to other factors in relation to other factors affect fermentation process