Effect of different spectral processing methods on the quantitative model of Astragalus polysaccharides detected by NIR spectroscopy

Quality control of Chinese herbal medicines requires reliable and rapid detection techniques, Many scholars have used NIR spectroscopy for research in medicine industry. The aim this study was to evaluate the feasibility combining machine learning predict adulteration content Astragalus polysaccharides. astragalus polysaccharide rice flour were mixed form adulterant with different concentration. Quantitative prediction models are built using support vector (SVM). Different pre-processing methods SG, SNV, WT, SG+SNV, SG+WT process spectra, then continuous projection algorithm (SPA) uninformative variable elimination (UVE) select characteristic wavelengths. SVM model based on SG+SNV+SPA is optimal an RPD 2.653, indicating that has good predictive ability. results showed technique could be quantitative analysis