Predicting the shelf life of Trachinotus ovatus during frozen storage using a back propagation (BP) neural network model

Abstract

The research aimed to create a shelf life prediction model for Trachinotus ovatus in different freezing temperatures by using back propagation (BP) neural network model. The pH, total volatile basic nitrogen (TVB-N), thiobarbituric acid (TBA), water retention (water holding capacity [WHC]; cooking loss), and sensory evaluation were measured under 266 K, 255 K, 243 K, 233 K, and 218 K temperatures. The results of TVB-N and water retention during 266 K, 255 K, 233 K, and 218 K conditions were selected to build a BP neural network model and verify the model at 243 K. Results indicated that low temperatures retarded the rise of pH, TVB-N, and TBA values, improving water retention capacity of Trachinotus ovatus. The BP neural network model had high regression coefficients (r2: 0.8642–0.9904), low mean square error (MES: 0.1658–1.7882), and relative error within 10% and could accurately predict the quality change of Trachinotus ovatus under the freezing temperatures of 266 K–218 K. Therefore, (BP) neural network model has great potential in predicting the shelf life of Trachinotus ovatus in frozen storage.

Keywords

• Trachinotus ovatus
• Shelf life prediction model
• BP neural Network
• Quality
• Water retention