论文部分内容阅读
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 a competitive self-organizing artificial neural network method for discriminating orange-red fingerprints of different cultivars. Methods The fingerprint of different cultivars was established by Fourier transform infrared spectroscopy and HPLC method. The self-organizing competitive artificial neural network was used for pattern recognition. Results The self-organizing competitive neural network model predicts the average accuracy of infrared spectra and HPLC fingerprint of the orange-red powder and extract by more than 91.67%. Conclusion Self-organizing competitive artificial neural network can be effectively used to identify the orange varieties.