Predicting Antioxidant Capacity of Whey Protein Hydrolysates Using Soft Computing Models

Whey proteins are considered as multi functional foods with several health benefits. Interest in utilizing whey proteins has increased substantially in recent years that are used as potential ingredients in several foods in the prevention of oxidation in fat-containing foodstuffs, cosmetics and pharmaceuticals. In this paper, predictive models based on soft computing paradigms including connectionist and neuro-fuzzy approaches as well as the conventional multiple regression technique are proposed to predict antioxidant capacity of whey protein hydrolysates. The performance of these models is compared with each other to assess their prediction potential. The results of this study revealed that the soft computing approach seemed to perform better than the conventional multiple regression technique. Also, among the two soft computing techniques used, the hybrid approach, i.e., neuro-fuzzy model performed the best.