Prediction of the freshness of horse mackerel (Trachurus japonicus) using E-nose, E-tongue, and colorimeter based on biochemical indexes analyzed during frozen storage of whole fish

Electronic nose (E-nose), electronic tongue (E-tongue) and colorimeter combined with data fusion strategy and different machine learning algorithms (artificial neural network, ANN; extreme gradient boosting, XGBoost; random forest regression, RFR; support vector regression, SVR) were applied to quantitatively assess and predict the freshness of horse mackerel (Trachurus japonicus) during the 90-day frozen storage. The results showed that the fusion data of the E-nose, E-tongue and colorimeter could contain more information (with a total variance contribution rate of 94.734 %) than that of the independent one. ANN, RFR and XGBoost showed good performance in predicting biochemical indexes with the RP2 (the square correlation coefficient of the Test set) ≥ 0.929, 0.936, 0.888, respectively, while SVR models showed a bad performance (RP2 ≤ 0.835). In addition, among the established quantitative models, the RFR model had the best prediction effect on K value (freshness index) with Rp2 of 0.936, ANN model had the highest fitting degree in predicting carbonyl content (protein oxidation degree) with Rp2 of 0.978, XGBoost model had the best performance in predicting the TBA value (lipid oxidation degree) with Rp2 of 0.994, RFR model was the best strategy for predicting Ca2+-ATPase activity (protein denaturation degree) with Rp2 of 0.969. The results demonstrated that the freshness of frozen fish can be effectively evaluated and predicted by the combination of electronic sensor fusion signals.