Effect of cold plasma processing on physicochemical and nutritional quality attributes of kiwifruit juice

Cold plasma treatment of kiwifruit juice was studied in the domain of 18-30 kV of voltage, 2-6 mm of juice depth, and 6-10 min of treatment time using the response surface methodology (RSM). The experimental design utilized was a central composite rotatable design. The effect of voltage, juice depth, and treatment time on the various responses, namely peroxidase activity, color, total phenolic content, ascorbic acid, total antioxidant activity, and total flavonoid content, was examined. While modeling, the artificial neural network (ANN) showed greater predictive capability than RSM as the coefficient of determination (R² ) value of responses was greater in the case of ANN (0.9538-0.9996) than in RSM (0.9041-0.9853). The mean square error value was also less in the case of ANN than in RSM. The ANN was coupled with a genetic algorithm (GA) for optimization. The optimum condition obtained from ANN-GA was 30 kV, 5 mm, and 6.7 min, respectively.