摘要
目前桑椹提取黄酮含量研究主要采用人工测量的方法,对其进行有效的预测较为困难。将主成分分析与BP神经网络结合建立科学、快速的桑椹黄酮提取含量预测模型。实测影响桑椹黄酮提取含量的4个因素组成数据样本,对该样本进行主成分分析,提取出影响桑椹黄酮提取含量的3个主成分,以这3个主成分数据作为BP神经网络模型的输入数据进行训练,用训练好的神经网络对桑椹黄酮提取含量进行预测。结果表明:该模型具有较高的预测精度,利用主成分分析和BP神经网络对桑椹黄酮提取含量进行预测以及检测是行之有效的。
At present,determination of flavonoids extraction content of mulberry is mostly done manually,which is difficult to be predicted. A scientific and rapid prediction model was created through combining principal component analysis with BP artificial neural network. Data of 4 factors influencing the flavonoids extraction content of mulberry was obtained through experiments,and 3principal components were extracted after principal component analysis of above data. BP artificial neural network was trained with above 3 principal components as input data,and then flavonoids extraction content of mulberry can be predicted through the trained BP artificial neural network.Experiment result shows that the prediction model has high prediction accuracy,so using principal component analysis and BP artificial neural network to predict flavonoids extraction content of mulberry is effective.
出处
《重庆理工大学学报（自然科学）》
CAS
2016年第6期96-101,共6页
Journal of Chongqing University of Technology：Natural Science
基金
福建省自然科学基金资助项目(2013J01377)
福建省教育厅A类项目(JA14087)
关键词
桑椹
黄酮
提取含量
主成分分析
BP神经网络
mulberry
flavonoid
extraction content
principal component analysis
BP artificial neural network
