Comparison of Artificial Neural Network and Response Surface Methodology Performance on Fermentation Parameters Optimization of Bioconversion of Cashew Apple Juice to Gluconic Acid

Abstract

Abstract The study examined the impact and interactions of cashew apple juice (CAJ) concentration, pH, NaNO3 concentration, inoculum size and time on gluconic acid (GA) production in a central composite design (CCD). The fermentation process and parameters involved were modeled and optimized using artificial neural network (ANN) and response surface methodology (RSM). The ANN model established the optimum levels as CAJ of 250 g/l, pH of 4.21, NaNO3 of 1.51 g/l, inoculum size of 2.87% volume and time of 24.41 h with an actual GA of 249.99 g/l. The optimum levels predicted by RSM model for the five independent variables were CAJ of 249 g/l, pH of 4.6, NaNO3 of 2.29 g/l, inoculum size of 3.95% volume, and time of 38.9 h with an actual GA of 246.34 g/l. The ANN model was superior to the RSM model in predicting GA production. The study demonstrated that CAJ could serve as the sole carbon source for GA production.