Machine learning-based optimization of alginate, guar gum, and pectin-based edible coatings for extended strawberry shelf life
Saklain Niam, Iftekhar Ahmad, Md Abu Rayhan, Sajid Mahmood, Parvej Hasan Jon, Md. Monir Ahmed
Abstract
Strawberries deteriorate rapidly under warm, humid conditions, and in many tropical settings, refrigerated logistics are limited. Edible biopolymer coatings offer a low-cost method to slow quality loss, yet multicomponent formulations behave nonlinearly. In this study, ten types of coatings were developed and evaluated for their effects on the physicochemical (weight loss, firmness, total soluble solids (TSS), pH, color change (ΔE)) and biochemical properties (total phenolic content (TPC), antioxidant property (DPPH) of strawberries over a 6-day storage period at 25 °C (298.15K). The optimal formulation, consisting of 0.56% (w/v) alginate, 0.33% (w/v) guar gum, and 2.11% (w/v) pectin, shows 7% weight loss, 586.34 N firmness, a pH of 4.08, and a ΔE of 4.32 according to our machine learning models. Additionally, two formulation families emerged from the optimization: pectin-rich coatings extended shelf-life with lower weight loss and restrained color change, while alginate-rich coatings maintained better firmness. All machine learning models (Random Forest (RF), Support Vector Regression (SVR), and eXtreme Gradient Boosting (XGB)) achieved high predictive accuracy (R 2 > 0.97). • Expended 33 data points into 132 using controlled random noise addition for data augmentation. • Machine learning models (SVR, XGB, RF) showed: R 2 > 97%. • The optimized coating had 0.56% alginate, 0.33% guar gum, and 2.11% pectin. • The optimal coating reduced 77.7% weight loss, and reduced 76% color change, improved 4.27x firmness and 4.65x TSS retention over 6 days storage.