Artificial neural network modeling of engineering properties and microbial stability of watermelon juice preserved with sustainable natural additives

Food quality preservation remains a major challenge globally, especially for highly labile fruit juices. This study was conducted to predict the engineering and microbial stability of bio-fortified watermelon juice, using linear regression analysis and an artificial neural network (ANN). In this study, watermelon juice was treated with the synthetic additive sodium benzoate and natural (sustainable) additives, including clove seed extract, cinnamon extract, bay leaf extract, ginger rhizome extract, honey, and lime juice. The results revealed that the natural additives significantly improved, the engineering properties, antioxidant ability, and antimicrobial potential of the treated juice samples ( p ≤ 0.05), with some of the multi-component hybridized treatment formulations (particularly Treatment T13), performing better than the synthetic additive. Particularly, the negative control unit exhibited the poorest DPPH, TPC, TFC, FRAP, Vitamin C, and antimicrobial values, while the T13 sample recorded the maximum values of 91.30% DPPH, 191.86 mg GAE/100 mL juice, 77.31 mg QE/100 mL juice, 10.37 μmol Fe 2+ /mL, and 17.82 mg/100 mL, respectively. Also, the T13 sample recorded the lowest bacterial and fungal counts of 3.29 ×10 4 and 2.17 ×10 3 CFU/mL after 10 storage days, respectively. The results of the linear regression analysis, microbial inhibition efficiency, and index comparison further affirmed that the sustainable additives can substantially improve the antioxidant and antimicrobial mechanisms of the treatment. The linear regression modeling elucidated the associations between treatment and storage time and the evaluated quality attributes; conversely, the ANN approach acted as a supplementary nonlinear strategy for anticipating selected bioprocess and microbial behavior. Basically, the findings of this research have highlighted that, combined natural additives and computer programming, have potential to improve antioxidant effects, microbial stability, and targeted biophysical characteristics of watermelon juice. These findings suggest preservation capacity without demonstrating definitive clinical or epidemiological benefits.