The maximal deoxynivalenol （DON） and fumonisins 131 ＋ B2 （FUM） contents in cereals are dictated by the European regulation 1126/2007. The direct measurement of these mycotoxins is a tedious and expensive process....The maximal deoxynivalenol （DON） and fumonisins 131 ＋ B2 （FUM） contents in cereals are dictated by the European regulation 1126/2007. The direct measurement of these mycotoxins is a tedious and expensive process. Our study is based on an alternative tool： near infrared spectroscopy. Different models were developed on 374 maize samples to predict their DON and FUM contents. Several parameters have been determined and used in a multivariate data analysis. Three models were developed： （1） a classification model based on Discriminant Factor Analysis （DFA）, （2） a linear model based on ANalysis of COVAriance （ANCOVA） and （3） a Partial Least Squares Discriminant Analysis model （PLS-DA）. Firstly, the performances of the DFA model for assessing DON and FUM risk were similar： 69 and 72% of the validation samples were respectively well classified. In the second part, the performances of the ANCOVA model for DON were higher than for FUM. The r2 was worth respectively 0.85 and 0.69. In the last part, the performances of the PLS-DA models were better for FUM than for DON. These results show that an evaluation of the mycotoxin risk is possible by analyzing selected kernel parameters measurable by secondary analytical such as near-infrared spectroscopy. Further work is needed to improve the models, adding more samples and using non linear approaches.展开更多
Consensus methods have presented promising tools for improving the reliability of quantitative models in near-infrared(NIR) spectroscopic analysis.A strategy for improving the performance of consensus methods in multi...Consensus methods have presented promising tools for improving the reliability of quantitative models in near-infrared(NIR) spectroscopic analysis.A strategy for improving the performance of consensus methods in multivariate calibration of NIR spectra is proposed.In the approach,a subset of non-collinear variables is generated using successive projections algorithm(SPA) for each variable in the reduced spectra by uninformative variables elimination(UVE).Then sub-models are built using the variable subsets and the calibration subsets determined by Monte Carlo(MC) re-sampling,and the sub-model that produces minimal error in cross validation is selected as a member model.With repetition of the MC re-sampling,a series of member models are built and a consensus model is achieved by averaging all the member models.Since member models are built with the best variable subset and the randomly selected calibration subset,both the quality and the diversity of the member models are insured for the consensus model.Two NIR spectral datasets of tobacco lamina are used to investigate the proposed method.The superiority of the method in both accuracy and reliability is demonstrated.展开更多
文摘The maximal deoxynivalenol （DON） and fumonisins 131 ＋ B2 （FUM） contents in cereals are dictated by the European regulation 1126/2007. The direct measurement of these mycotoxins is a tedious and expensive process. Our study is based on an alternative tool： near infrared spectroscopy. Different models were developed on 374 maize samples to predict their DON and FUM contents. Several parameters have been determined and used in a multivariate data analysis. Three models were developed： （1） a classification model based on Discriminant Factor Analysis （DFA）, （2） a linear model based on ANalysis of COVAriance （ANCOVA） and （3） a Partial Least Squares Discriminant Analysis model （PLS-DA）. Firstly, the performances of the DFA model for assessing DON and FUM risk were similar： 69 and 72% of the validation samples were respectively well classified. In the second part, the performances of the ANCOVA model for DON were higher than for FUM. The r2 was worth respectively 0.85 and 0.69. In the last part, the performances of the PLS-DA models were better for FUM than for DON. These results show that an evaluation of the mycotoxin risk is possible by analyzing selected kernel parameters measurable by secondary analytical such as near-infrared spectroscopy. Further work is needed to improve the models, adding more samples and using non linear approaches.
基金supported by the National Natural Science Foundation of China (20835002)
文摘Consensus methods have presented promising tools for improving the reliability of quantitative models in near-infrared(NIR) spectroscopic analysis.A strategy for improving the performance of consensus methods in multivariate calibration of NIR spectra is proposed.In the approach,a subset of non-collinear variables is generated using successive projections algorithm(SPA) for each variable in the reduced spectra by uninformative variables elimination(UVE).Then sub-models are built using the variable subsets and the calibration subsets determined by Monte Carlo(MC) re-sampling,and the sub-model that produces minimal error in cross validation is selected as a member model.With repetition of the MC re-sampling,a series of member models are built and a consensus model is achieved by averaging all the member models.Since member models are built with the best variable subset and the randomly selected calibration subset,both the quality and the diversity of the member models are insured for the consensus model.Two NIR spectral datasets of tobacco lamina are used to investigate the proposed method.The superiority of the method in both accuracy and reliability is demonstrated.
