Prediction of Mass Transfer Kinetics During Osmotically Pretreated Heat Sensitive Products: Experimental Investigation, ANOVA and ANFIS Modeling

ABSTRACT The primary objective of this research is to develop and validate a robust Adaptive Neuro‐Fuzzy Inference System (ANFIS) model for predicting mass transfer characteristics in eggplant during osmotic dehydration. This study introduces the first ANFIS model adapted to eggplant's lignified cellular structure, advancing beyond prior work on zucchini 9. Experiments used sucrose concentrations (40%–60%), temperatures (25°C–35°C), and immersion times (0–120 min). Key findings include: ANFIS reduced water loss (WL) prediction error by 91% versus ANOVA; sensitivity analysis identified temperature as the dominant parameter (Sobol index = 0.62); and optimized conditions (60% Brix, 35°C, 120 min) achieved a WL/solid gain (SG) ratio of 13.82. ANFIS outperformed ANN/SVM by 7.7% in accuracy, demonstrating its superiority for industrial dryer design.