Notice of Retraction Combination of ^1H NMR and Chemometrics for Discrimination of Rapeseed Oils

1 H NMR spectroscopy was utilized to distinguish the brands of rapeseed oils. As there are more than four hundreds of NMR variables, uninformative variables should be eliminated to improve model's discrimination ability and save the calculation time. A hybrid variable selection which is combined with uninformation variable elimination (UVE) and successive projections algorithm (SPA), was used to achieve this objective. 77 effective variables were selected from the full-spectrum variables. They were inputted into least-square support vector machine (LS-SVM) to establish the discrimination model. A good result of 92.5% correct answer rate was obtained. It is improved compared to the result of the full spectrum- SPA-LS-SVM model. It is proved that it is necessary to do UVE before SPA. As a conclusion, the performance of UVE-SPALS-SVM model shows that 1 H NMR spectroscopy is a feasible way to distinguish rapeseed oils fast and accurately.