Tang Yan-feng, Zhang Zhuo-yong, Fan Guo-qiang
Department of Chemistry Resources Environment and GIS Key Lab of Beijing, Capital Normal University, Beijing, China.
Guang Pu Xue Yu Guang Pu Fen Xi. 2005 Apr;25(4):521-4.
Takagi-Sugeno fuzzy system is composed of several back-propagation neural networks (BP-NNs), and has some fuzzy logic properties. In this paper, the Takagi-Sugeno fuzzy logic system is applied to identifying official and unofficial rhubarb samples based on their infrared spectra. The effects of the number of hidden neurons and the momentum parameters on the prediction were investigated. The results obtained by using Takagi-Sugeno fuzzy system were better than those by commonly used BP-networks. With a proper network training parameter, 100% correctness can be obtained by using Takagi-Sugeno fuzzy system. This method is more accurate than the common methods, and is more scientific than traditional methods. So it is applied to identifying rhubarb easily and rapidly.
高木-关野模糊系统由多个反向传播神经网络（BP-NN）组成，并具有一些模糊逻辑属性。本文将高木-关野模糊逻辑系统应用于基于红外光谱识别正品和非正品大黄样本。研究了隐藏神经元数量和动量参数对预测的影响。使用高木-关野模糊系统获得的结果优于常用的BP网络。通过适当的网络训练参数，使用高木-关野模糊系统可获得100%的正确率。该方法比常用方法更准确，比传统方法更科学。因此，它可轻松、快速地应用于大黄的识别。
