[Quantitative prediction of active constituents in rhubarb by near infrared spectroscopy and radial basis function neural networks].

Near infrared spectroscopy (NIRS) and artificial neural networks were used for the quantitative prediction of four active constituents in rhubarb: anthraquinones, anthraquinone glucosides, stilbene glucosides, Tannins and related compounds. The near infrared spectra of the samples were acquired in 1100-2500 nm from powdered rhubarb samples. Four calibration models using radial basis function neural networks (RBFNN) were set up to correlate the spectra with the values determined by HPLC. RMSECVs of the models for the constituents studied were 2.572, 0.442, 2.794 and 9.438, respectively. RMSEPs for the were 4.598, 8.657, 0.4586, and 5.106, respectively. The method is fast, and satisfactory results were obtained. The proposed method can be used for determining the active constituents in Chinese herbal medicine.