Fractional factorial design and machine learning reveal key formulation factors in chickpea-based (Cicer arietinum) cheese analogues

The growing demand for plant-based cheese analogues (PBCA) reflects increasing consumer interest in dairy-free products with desirable texture and sensory attributes. This study applied a hybrid optimization framework combining fractional factorial design (FFD) and artificial neural network (ANN) modeling to identify key formulation factors influencing chickpea-based (Cicer arietinum) cheese analogues. Initially, several plant-based milks (mung bean, whole mung bean, black-eyed pea, vigna mungo, chickpea, and soybean) were evaluated with different coagulants to determine milk-clotting efficiency. Chickpea milk exhibited superior coagulation with papain, and was subsequently selected for formulation optimization. The effects of water, vegetable oil, starch, papain, and carrageenan on textural properties were analyzed. The FFD results showed that starch significantly increased hardness, gumminess, and chewiness, while carrageenan enhanced cohesiveness and structural stability (p < 0.05). The ANN model demonstrated high predictive accuracy (R² = 0.999 for hardness, 0.913 for chewiness, 1.000 for cohesiveness), capturing nonlinear interactions beyond the quadratic limitations of FFD. The integration of FFD and ANN effectively minimized experimental variability and revealed starch and carrageenan as critical determinants of plant-based cheese analogue (PBCA) texture and mouthfeel. The integrated modeling strategy reduces experimental trials and material use, supporting resource-efficient formulation development. This keeps sustainability without invoking policy goals.

Text 125871.pdf - Published Version Restricted to Repository staff only Download (1MB) Official URL or Download Paper: https://link.springer.com/10.1007/s11694-026-04430...

Actions (login required)

View Item