基于自组织竞争型神经网络的化橘红指纹图谱研究

Research of Exocarpium Citri Grandis Fingerprint Based on Self-organization Competitive Neural Network

目的建立不同品种化橘红指纹图谱的自组织竞争型人工神经网络判别方法。方法采用傅里叶红外光谱法及HPLC法建立不同品种化橘红的指纹图谱,采用自组织竞争型人工神经网络进行模式识别。结果自组织竞争型神经网络模型对化橘红粉末、提取物红外及HPLC指纹图谱预测平均准确率达91.67%以上。结论自组织竞争型人工神经网络可有效用于化橘红品种的识别。

Objective To establish the self-organization competitive neural network method for identifying the species by the fingerprints of different kinds of Exocarpium Citri Grandis （Cirrus grandis Peel）. Methods The fingerprints of different species of Exocarpium Citri Grandis were established by fourier transform infrared spectroscopy and high performance liquid chromatogram （HPLC）. And then we used the self-organization competitive neural network model to distinguish the different species of Exocarpium Citri Grandis. Results The self-organization competitive neural net- work for distinguish Exocarpium Citri Grandis of different species was established, and the rate of accuracy for infrared and HPLC fingerprint of the powder and the extract was more than 91.67 % in average. Conclusion The self-organizing competitiveneural network can he used for distinguishing different species of Exocarpium Citri Grandis by their fingerprints.