ESTABLISHMENT OF THE ARRHENIUS MODEL AND THE RADIAL BASIS FUNCTION NEURAL NETWORK (RBFNN) MODEL TO PREDICT QUALITY OF THAWED SHRIMP (SOLENOCERA MELANTHO) STORED AT DIFFERENT TEMPERATURES

摘要

Changes in quality of thawed shrimp (Solenocera melantho) stored at −3, 0, 3 and 6C were determined and modeled by the Arrhenius and the radial basis function neural network (RBFNN) models, based on total volatile base nitrogen , total aerobic counts, K value, hypoxanthine, pH, electrical conductivity (EC) and sensory assessment. A significant inhibition of spoilage was found in shrimp stored at −3C compared to those stored at 0, 3 and 6C. The prediction accuracy of the Arrhenius model was satisfactory for indicators of EC and pH (relative errors within ± 10%), and relative errors for other indicators surpassed ± 10% on some days; the relative errors of RBFNN model were all within ± 2% for each indicator. The RBFNN model was more suitable for predicting the quality of thawed shrimp stored at −3 to 6C.