Computational modeling for the enhancement of thermosonicated Sohphie (Myrica esculenta) fruit juice quality using artificial neural networks (ANN) coupled with a genetic algorithm

Abstract

This study investigated the impact of thermosonication on enhancing the nutritional characteristics of juice derived from Sohphie (Myrica esculenta) fruits. This investigation introduces an innovative approach utilizing artificial neural networks (ANNs) for the multifaceted optimization of the juice extraction process. Specifically, we focused on determining the most effective extraction parameters for thermosonication, including amplitude (30%, 40%, and 50%), treatment time (15, 30, 45, and 60 min) and temperature (30 °C, 40 °C, and 50 °C). The primary objective of this approach was to augment the nutritional and microbiological properties of Sohphie juice by improving its quality attributes such as ascorbic acid (AA) content, anti-oxidant activity (AOA), total anthocyanin content (TAC), total carotenoid content (TCC), total flavonoid content (TFC), total phenolic content (TPC), total viable count (TVC), and yeast and mould count (YMC). The maximum levels of AA (58.74 ± 3.56 mg/100 mL), AOA (66.11% ± 3.92%), TAC (48.50 ± 4.57 μg mL⁻¹), TCC (133.60 ± 5.17 βCE μg mL⁻¹), TFC (55.49 ± 3.86 mg quercetin equivalents (QE) per mL), TPC (78.94 ± 4.84 mg gallic acid equivalents (GAE) per mL), TVC (2.44 ± 0.23 log CFU mL⁻¹) and YMC (1.01 ± 0.11 log CFU mL⁻¹) were obtained in thermosonicated Sohphie juices (TSSJ) under optimal conditions. This study highlights that artificial neural networks (ANNs) coupled with a genetic algorithm (GA) are a beneficial resource for forecasting the extraction efficiency of Sohphie fruit juice (SJ) and suggests that employing thermosonication as a preservation method for SJ can potentially replace traditional thermal pasteurization. This strategy has the potential to reduce or prevent quality deterioration while enhancing the functionality of the juice.