Classification of Sulfadimidine and Sulfapyridine in Duck Meat by Surface Enhanced Raman Spectroscopy Combined with Principal Component Analysis and Support Vector Machine

In order to meet the growing demand for food safety, the rapid classification of sulfadimidine and sulfapyridine in duck meat based on surface-enhanced Raman spectroscopy (SERS) was investigated using gold nanoparticles (Au NPs) and NaCl solution as the active substrate and active agent, respectivel...

Full description

Saved in:
Bibliographic Details
Published in:Analytical letters 2020-07, Vol.53 (10), p.1513-1524
Main Authors: Ning, Xu, Mu-Hua, Liu, Hai-Chao, Yuan, Shuang-Gen, Huang, Xiao, Wang, Jin-Hui, Zhao, Jian, Chen, Ting, Wang, Wei, Hu, Yi-Xin, Song
Format: Article
Language:English
Subjects:
Classification
Duck meat
Gold
gold nanoparticles
Iterative methods
Meat
Nanoparticles
Optimization
Pretreatment
principal component analysis
Principal components analysis
Radial basis function
Raman spectroscopy
Sensitivity
Spectrum analysis
Substrates
sulfadimidine
sulfapyridine
support vector machine
Support vector machines
surface-enhanced Raman spectroscopy
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:In order to meet the growing demand for food safety, the rapid classification of sulfadimidine and sulfapyridine in duck meat based on surface-enhanced Raman spectroscopy (SERS) was investigated using gold nanoparticles (Au NPs) and NaCl solution as the active substrate and active agent, respectively. The results showed that the Raman characteristic peaks at 557 and 573 cm −1 were used to determine whether sulfadimidine and sulfapyridine remained in duck meat. A single factor experimental method was used to optimize the conditions. Three pretreatment methods were employed to optimize the raw measurements. According to the classification accuracies, adaptive iterative reweighted penalty least squares and second derivative were selected to be the optimal pretreatment method. Next, principal component analysis (PCA) was performed to extract the characteristic variables. The first four score values of PCA were selected to be the input values of support vector machine (SVM) classification model, which classified duck meat samples into four categories. Nu-support vector classification and radial basis function were selected to be the employed SVM classification model and the Kernel type, respectively. The gamma and nu values were 0.25 and 0.255, respectively. The classification accuracy for the test set was 90.44%. The sensitivities and specificities of test set were calculated, and the highest sensitivity and specificity values were 96.97% and 100%, respectively. The experimental results showed that this classification model provides favorable classification results. Therefore, the adopted method may be used for the rapid classification of sulfadimidine and sulfapyridine residues in duck meat.
ISSN:0003-2719
1532-236X
DOI:10.1080/00032719.2019.1710524