Application of Artificial Neural Network to Predict K-Value, Inosine Mono-Phosphate, and Hypoxanthine Concentrations of Grass Carp (Ctenopharyngodon idellus) Fillets During Storage

作者：Qin, Na; Li, Dapeng; Li, Qian; Li, Dongping; Liu, Xiaochang; Luo, Yongkang^*

来源：International Journal of Food Properties, 2016, 19(12): 2693-2706.

DOI：10.1080/10942912.2016.1141425

• Grass carp fillets
• Inosine mono-phosphate
• Artificial neural network
• Hypoxanthine
• K-value
• Ctenopharyngodon idellus

K-value, inosine mono-phosphate, and hypoxanthine concentrations of grass carp (Ctenopharyngodon idellus) fillets were determined during storage at 273, 277, 281, 288, and 293 K. Simultaneously, a feed-forward artificial neural network was developed to predict these changes in grass carp fillets during storage, and a comparative study on K-value prediction between the artificial neural network and Arrhenius model was also performed. The results showed that the K-value and hypoxanthine concentrations increased with storage time, while inosine mono-phosphate reached a peak and then decreased with time. The artificial neural network was successful in predicting changes in the K-value, inosine mono-phosphate, and hypoxanthine concentrations throughout storage, and it was even more effective in predicting K-value with lower relative errors than the Arrhenius model. The high regression coefficient (R-2) and low mean squared error indicated that the artificial neural network could be a potential tool in modeling changes in K-value, inosine mono-phosphate, and hypoxanthine concentrations of grass carp fillets within 273-293 K.

• 出版日期2016
• 单位中国农业大学