کد مقاله | کد نشریه | سال انتشار | مقاله انگلیسی | نسخه تمام متن |
---|---|---|---|---|
6391634 | 1628420 | 2014 | 5 صفحه PDF | دانلود رایگان |
- A portable near infrared spectrometer was used for total bacteria in fishes.
- The NIR model exploited genetic algorithm and artificial neural network.
- The NIR data were pretreated by wavelet transform to diminish interference.
There is an emerging trend of non-destructive and onsite analysis of microbial contaminations for better food safety. A new strategy for determination of total bacterial in fish products (flounder fillets) was established using a portable near infrared spectrometer. Results revealed that the pretreatment of near infrared spectrum by the wavelet transform could significantly improve the accuracy and precision of the analysis. In comparison to usually exploited partial least squares regression (PLS), a combination of genetic algorithm (GA) and back-propagation artificial neural network (BP-ANN) exhibited much better efficiency, and the correlation coefficient (R) and root mean square error (RMSE) of the prediction model were calculated as 0.985 and 0.095, respectively, and validated as 0.966 and 0.083, respectively. These results allowed us to suggest a promising potential of the established technique for non-destructive and onsite monitoring of total bacteria in fishery products.
Journal: Food Control - Volume 42, August 2014, Pages 18-22