Prediction of insoluble dietary fiber content in moso bamboo shoot using near-infrared spectroscopy

Near-Infrared spectroscopy (NIRS) was used to determine the insoluble dietary fiber (IDF) content of moso bamboo shoot. A total of 120 samples were separated into two groups, as calibration set and prediction set. The IDF content of samples ranged from 0.046% to 0.209% were measured by AOAC 991.43 method. Spectral data from 400 nm to 1000 nm were collected using the samples. Three types of preprocessing methods including standard normal variate (SNV), multiplicative scattering correction (MSC), and Savitzky-Golay first derivative transformation (S. Golay 1st-Der) were used before the spectral data were analyzed. Partial least squares (PLS) analysis was performed on the calibration set, and calibration models were developed using the different preprocessed spectral data. The root mean square error of calibration (RMSEC) and correlation coefficient (r) of the models were calculated to select the best model. The MSC-PLS model with r of 0.950 and RMSEC of 0.008 showed an excellent prediction performance. Thus, MSC model was selected as the best model for prediction of IDF content. The over all results show that NIRS technique coupled with PLS could be applied as an alternative fast and accurate method to determine IDF content in moso bamboo shoot.