Abstract:It was difficult to determined the content of whole potato powder and flour in potato noodles by chemical method. The objective of the study was to establish prediction model for rapid determination of potato powder composition in noodles by NIR spectroscopy and stoichiometry. The method was that 236 potato noodles studied by near infrared spectroscopy and stoichiometry were used to establish prediction model. The result indicated that the prediction model based on wave number of 9403.6~5446.2 cm-1 and spectral pretreatment by minimum-maximum normalization showed an excellent prediction accuracy. The external validation determinant (R2val) was 0.9775, the root mean square error of prediction (RMSEP) was 1.28%, the slope was 0.95, and the relative prediction deviation (RPD) was 4.74. The results indicated that near infrared spectroscopy can be used as an efficient way to detect the potato powder content in potato noodles.
