Quantitative analysis using NIR by building principal component-multiple linear regression-BP algorithm

Near infrared reflectance spectroscopy (NIRS) appears to be a rapid and convenient non-destructive technique that can measure the quality and compositional attributes of many substances. This paper assesses the ability of NIR reflectance spectroscopy to estimate the pH values of bayberry juice. Spectra were collected from 76 juice samples and data was expressed as absorbance, the logarithm of the reciprocal of reflectance (log1/R). The absorbance data was subsequently compressed using wavelet transformation. Three models to predict the acidity in bayberry juice were constructed. A prediction model based on principle component analysis-multiple linear regression-back propagation (PCA-NMR-BP) was found to be superior (r=0.934, RMSEP=0.263) to models based on PCA-BP and MLR-BP.