基于随机森林的原位质谱法快速鉴别铁棍山药真伪
Rapid identification of the authenticity of iron rod yam by in-situ mass spectrometry based on random forest algorithm
钟恒艳 1陈春 1欧阳永中 2周林 3郭伟清4
作者信息
- 1. 佛山大学,广东 佛山 528000
- 2. 佛山大学,广东 佛山 528000;佛山职业技术学院,广东 佛山 528137
- 3. 广东一方制药有限公司,广东 佛山 528244
- 4. 佛山大学,广东 佛山 528000;有研(广东)新材料技术研究院,广东 佛山 528000
- 折叠
摘要
[目的]建立一种快速无损的铁棍山药真伪鉴别方法.[方法]在常温常压环境下,运用大气压化学电离质谱技术对不同产地的铁棍山药(TG)和非铁棍山药(FTG)化学成分进行检测,每种铁棍山药样品和非铁棍山药样品各采集200组数据,共获得3 600个质谱数据点,利用主成分分析(PCA)和随机森林(RF)算法对获得的一级质谱数据进行模式识别分析,建立铁棍山药与非铁棍山药真伪鉴别模型.[结果]通过HS-APCI-MS获得的铁棍山药样品与非铁棍山药样品的一级质谱图差异明显,根据主成分累计方差贡献图,前7个主成分的累计方差贡献率为85.63%(≥85%);当决策树数量为25时,训练集、检测集准确率均达到100%,所建立的铁棍山药原位质谱分析结合RF算法对铁棍山药的鉴别效果显著,RF的分类效果优于PCA.[结论]应用原位质谱分析技术结合RF算法可以快速无损鉴别铁棍山药真伪.
Abstract
[Objective]To establish a fast and nondestructive analysis method for identifying iron rod yam.[Methods]Atmospheric pressure chemical ionization mass spectrometry(APCI-MS)was employed to detect the chemical constituents of iron rod yam(TG)and non-iron rod yam(FTG)from different origins under ambient temperature and pressure,With 200 sets of data collected from each type of TG and FTG,and a total of 3 600 mass spectrometry data points were obtained.Subsequently,the initial level of the mass spectrometry data obtained was analyzed using Principal Component Analysis(PCA)and the random forest(RF)algorithm.Pattern recognition analysis established a model to differentiate between TG and FTG based on their chemical compositions.[Results]The difference between the first-level mass spectra obtained by HS-APCI-MS was obvious between TG samples and FTG samples.The cumulative variance contribution plot of the principal components showed that the first seven principal components accounted for 85.63%(≥85%)of the variance.The accuracy of the training set and detection set reached 100%when the number of decision trees was 25.HS-APCI-MS combined with RF algorithm had a significant identification effect on TG,and the classification effect of RF was superior to that of PCA.[Conclusion]Atmospheric pressure chemical ionization mass spectrometry,combined with the RF algorithm,can rapidly and non-destructively identify TG and FTG,providing a new technical method for authenticating TG.
关键词
铁棍山药/大气压化学电离源/主成分分析/随机森林算法/鉴别Key words
iron rod yam/atmospheric chemical ionization source/principal component analysis/random forest algorithm/identification引用本文复制引用
出版年
2024
国家科技期刊平台
NSTL
万方数据