Search

BACKGROUND Freshness is an important quality of squid with respect to determining the market price. The methods of evaluation of freshness fail to be widely used as a result of the lack of rapidity and quantitation. In the present study, a rapid and non‐destructive quantification of squid freshness by Fourier transform infrared spectroscopy (FTIR) spectra combined with chemometric techniques was performed. RESULTS: The relatively linear content change of trimethylamine (TMA‐N) and dimethylamine (DMA‐N) of squid during storage at 4 °C indicated their feasibility as a freshness indicator, as also confirmed by sensory evaluation. The spectral changes were mainly caused by the degradation of proteins and the production of amines by two‐dimensional infrared correlation spectroscopy, among which TMA‐N, DMA‐N and putrescine were the main amines. The successive projections algorithm (SPA) was employed to select the sensitive wavenumbers to freshness for modeling prediction including partial least‐squares regression, support vector regression (SVR) and back‐propagation artificial neural network. Generally, the SPA‐SVR model of the selected characteristic wavenumber showed a higher prediction accuracy for DMA‐N (R2P = 0.951; RMSEP = 0.218), whereas both SPA‐SVR (R2P = 0.929; RMSEP = 2.602) and Full‐SVR (R2P = 0.941; RMSEP = 2.492) models had a higher predictive ability of TMA‐N. CONCLUSION: The results of the present study demonstrate that FTIR spectroscopy coupled with multivariate calibration shows significant potential for the prediction of freshness in squid. © 2021 Society of Chemical Industry [ABSTRACT FROM AUTHOR]