摘要
为了实现秦艽药品样本产地划分,将近红外光谱技术(Near infrared spectroscopy,NIRs)与深度森林(Deep Forest,DF)算法结合在甘肃不同产地的秦艽药品样本数据上构建定性分析模型.研究结果表明,模型对秦艽测试集样本的鉴定准确率为93%,F-Score可达0.98.说明近红外光谱与深度森林算法结合能够有效鉴定甘肃不同产地的秦艽样品,方法测试便捷效率高、仪器成本低,该方法能够为其他植物类中药的定性分析工作提供一定的参考.
In order to realize the division of Gentiana macrophylla drug sample origin,near infrared spectroscopy(NIRS)and deep forest(DF)algorithm are combined to build a qualitative analysis model on the Gentiana macrophylla drug sample data from different habitats in Gansu Province.The study results show that the identification accuracy of the model for Gentiana macrophylla test set samples is 93%,and F-Score could reach 0.98.It shows that the combination of near-infrared spectroscopy and deep forest algorithm can effectively identify Gentiana macrophylla samples from different habitats in Gansu Province.The method is convenient and efficient,and the instrument cost is low.This method can provide some reference for the qualitative analysis of other herbal medicines.
作者
丁跃武
杨友
陈方方
李四海
Ding Yue-wu;Yang You;Chen Fang-fang;Li Si-hai(College of Information Technology,Gansu University of Chinese Medicine,Lanzhou 730000,Gansu Province,China)
出处
《科学与信息化》
2023年第1期34-36,共3页
Technology and Information
基金
甘肃省科技计划项目,项目名称:近红外光谱的正则化特征选择算法研究,项目编号:21JR1RA272。
关键词
深度森林
近红外光谱
秦艽
定性分析
deep forest
near infrared spectroscopy
Gentiana macrophylla
qualitative analysis