NIRS Determination of Total Flavonoids in Panax Notoginseng with Consensus Radial Basis Function Neural Network

Consensus strategy was applied to radial basis function neural network and used in NIRS determination of total flavonoids in panax notoginseng. Firstly, the spectra were preprocessed using discrete wavelet transform for noise filtering and data compression. Then, consensus radial basis function neural network was used for establishing the calibration model. It was shown by the results that the consensus models were more stable and accurate than the conventional regression models.