Artificial Neural Network–Genetic Algorithm to Optimize Yin Rice Inoculation Fermentation Conditions for Improving Physico-chemical Characteristics

抄録

In this research, a nonlinear model describing the relationship between the inoculation fermentation parameters and the quality of yin rice were investigated based on artificial neural network and genetic algorithm (ANN-GA) model. The ANN-GA model had excellent potential for predicting the viscosity property of yin rice, and fermentation parameters were optimized by using genetic algorithm. Through ANN-GA model, the optimized inoculation fermentation parameters were: 0.05 % lactic acid bacteria, 0.05 % Saccharomyces cerevisiae, 0.2 % Rhizopus oryzae, then fermenting for 48 h at 25 °C. The results were further validated by experiments. Moreover, it revealed that inoculation fermentation not only effectively improved physico-chemical characteristics of yin rice, but also shorten period of fermentation about 14 days compared to the natural fermentation. These results indicated that the accuracy and reliable of fermentation parameters optimized by ANN-GA model.