基于机器学习鉴别牛黄类药材红外光谱的研究

doi:10.19803/j.1672-8629.20210405

中国药物警戒 ›› 2023, Vol. 20 ›› Issue (2): 140-145.
DOI: 10.19803/j.1672-8629.20210405

基于机器学习鉴别牛黄类药材红外光谱的研究

石岩¹, 王晓伟^2Δ, 魏锋^1*, 马双成^1#

¹中国食品药品检定研究院中药民族药检定所,北京 102629;
²河南省食品药品检验所,国家药品监督管理局中药材及饮片质量控制重点实验室,河南郑州 450018

收稿日期:2021-04-26 出版日期:2023-02-15 发布日期:2023-02-17
通讯作者: ^*魏锋,男,博士,研究员,中药质量评价与控制。E-mail: weifeng@nifdc.org.cn;^#为共同通信作者。
作者简介:石岩,男,博士,研究员,中药质量评价与控制。^Δ为并列第一作者。
基金资助:
国家重点研发计划（2019YFC1711500）; 国家科技重大专项重大新药创制（2018ZX09735-006）

Succession medicinal substances of calculus bovis with infrared spectroscopy coupled with machine learning methods

SHI Yan¹, WANG Xiaowei^2Δ, WEI Feng^1*, MA Shuangcheng^1#

¹National Institutes for Food and Drug Control, Beijing 102629, China;
²Henan Institute for Food and Drug Control, NMPA Key Laboratory for Quality Control of Traditional Chinese Medicine (Chinese Materia Medica and prepared slices), Zhengzhou Henan 450018, China

Received:2021-04-26 Online:2023-02-15 Published:2023-02-17

摘要/Abstract

摘要： 目的使用机器学习领域内的自组织映射神经网络对牛黄、体外培育牛黄和人工牛黄样品红外光谱数据进行区分和鉴别。方法将样品的红外光谱数据进行autoscale处理,然后分别进行非监督分析和监督分析。借助遗传算法研究和探讨最合适的模型类型及相关参数。选取的模型为XY-Fused神经网络,神经元数量为25（5×5）,epoch为1 000,训练方式为batch,神经元网络选择为有边界形式。然后将43批样品随机选取其中9批样品作为验证集样品,其余34批样品作为校正集用于模型的训练和学习。结果所建立的XY-Fused神经网络模型3次平行训练和学习的交叉验证正确率分别为94.1%、94.1%、94.1%,平均为94.1%;验证集验证正确率分别为100%、83.3%、100%,平均为94.4%。结论建立的XY-Fused神经网络适用于牛黄类药材的红外光谱数据分析和鉴别,对牛黄类药材的研究具有一定的参考价值。

关键词: 牛黄, 红外光谱, 机器学习, 数据分析, 鉴别

Abstract: Objective To explore a method based on machine learning for distinguishing succession medicinal substances of calculus bovis with infrared spectroscopy. Methods After being preprocessed using an autoscale method, data on infrared spectra was analyzed with unsupervised analysis methods and supervised ones respectively. With the help of a genetic algorithm, the most suitable model type and related parameters were studied. The most suitable model type was the XY-Fused network with 25(5×5) as the number of neurons and 1000 set as the epoch. The model with boundaries was trained with the batch pattern. Nine samples were selected randomly among the 43 samples as the test set. The rest were used as the calibration set. Results Accuracies of cross validation were 94.1%, 94.1%, 94.1%, respectively, and the average was 94.1%. Prediction accuracies for the test set were 100%, 83.3%, 100%, respectively, and the average was 94.4%. Conclusion The model established here can be used for distinguishing succession medicinal substances of cow-bezoar with infrared spectroscopy, and is of referential value for studies on succession medicinal substances of calculus bovis.

Key words: calculus bovis, infrared spectrum, machine learning, data analysis, identification

中图分类号:

R991
R994.11

石岩, 王晓伟, 魏锋, 马双成. 基于机器学习鉴别牛黄类药材红外光谱的研究[J]. 中国药物警戒, 2023, 20(2): 140-145.

SHI Yan, WANG Xiaowei, WEI Feng, MA Shuangcheng. Succession medicinal substances of calculus bovis with infrared spectroscopy coupled with machine learning methods[J]. Chinese Journal of Pharmacovigilance, 2023, 20(2): 140-145.

参考文献

[1] Chinese Pharmacopoeia Commission.Pharmacopoei of the People's Repablic of China, VolumI(中华人民共和国药典,一部)[M]. Beijing: China Medical Science Press, 2020.
[2] QIAO X, YE M, PAN DL, et al.Differentiation of various traditional Chinese medicines derived from animal bile and gallstone: Simultaneous determination of bile acids by liquid chromatography coupled with triple quadrupole mass spectrometry[J]. Journal of Chromatography A, 2011, 1218: 107-117.
[3] SHI Y, XIONG J, WEI F, et al.Study of characteristic ions in frequently-used traditional Chinese medicinal materials derived from animal bile and its application[J]. China Journal of Chinese Materia Medica(中国中药杂志), 2018, 43(4): 651-658.
[4] XIONG J, ZHENG TJ, SHI Y, et al.Analysis of the fingerprint profile of bioactive constituents of traditional Chinese medicinal materials derived from animal bile using the HPLC-ELSD and chemometric methods: An application of a reference scaleplate[J]. Journal of Pharmaceutical and Biomedical Analysis, 2019, 179: 50-56.
[5] LIU WX, CHENG XL, GUO XH, et al.Identification of Calculus Bovis and its mixed varieties by ultra-high-performance liquid chromatography/quadrupole time-of-flight mass spectrometry (UHPLC-Q/TOF-MS) combined with the principal component analysis (PCA) method[J]. Journal of Pharmaceutical and Biomedical Analysis, 2020, 179: 112979.
[6] KOHONEN T.The self-organizing map[J]. Proceedings of the IEEE, 1990, 78(9): 1464-1480.
[7] MELSSEN W, WEHRENS R, BUYDENS L.Supervised kohonen networks for classification problems[J]. Chemometrics and Intelligent Laboratory Systems, 2006, 83: 99-113.
[8] BALLABIO D, VASIGHI M.A MATLAB toolbox for self organizing maps and supervised neural network learning strategies[J]. Chemometrics and Intelligent Laboratory Systems, 2012, 118: 24-32.
[9] ABE T, AKAZAWA Y, TOYODA A, et al.Batch-learning self-organizing map identifies horizontal gene transfer candidates and their origins in entire genomes[J]. Frontiers in Microbiology, 2020, 11: 1486.
[10] GARDNER W, MALIKI R, CUTTS SM, et al.Self-organizing map and relational perspective mapping for the accurate visualization of high-dimensional hyperspectral data[J]. Analytical Chemistry, 2020, 92(15): 10450-10459.
[11] JAMMOUSSI I, NASR MB.A hybrid method based on extreme learning machine and self organizing map for pattern classification[J]. Computational Intelligence and Neuroscience, 2020, 2020(pt.2): 2918276.
[12] DETORAKIS G, CHAILLET A, ROUGIER NP.Stability analysis of a neural field self-organizing map[J]. Journal of Mathematical Neuroscience, 2020, 10: 20.
[13] GALVAN D, EFFTING L, CREMASCO H, et al.The spread of the COVID-19 outbreak in Brazil: an overview by kohonen self-organizing map networks[J]. Medicina, 2021, 57: 235.
[14] SUN L, MA X, JIN HY, et al.Geographical origin differentiation of Chinese Angelica by specific metal element fingerprinting and risk assessment[J]. Environmental Science and Pollution Research, 2020, 27(2): 1-13.
[15] GARDNER W, CUTTS SM, PHILLIPS DR, et al.Understanding mass spectrometry images: complexity to clarity with machine learning[J]. Biopolymers, 2020, 112(4): e23400.
[16] KATOCH S, CHAUHAN SS, KUMAR V.A review on genetic algorithm: past, present, and future[J]. Multimedia Tools and Applications, 2021, 80: 8091-8126.
[17] LLORELLA FR, PATOW G, AZORÍN JM. Convolutional neural networks and genetic algorithm for visual imagery classification[J]. Physical and Engineering Sciences in Medicine, 2020, 43(3): 973-983.
[18] CHEN Z, HUANG A, QIANG X.Improved neural networks based on genetic algorithm for pulse recognition[J]. Computational Biology and Chemistry, 2020, 88: 107315.
[19] ZHANG P, WEST NP, CHEN PY, et al.Selection of microbial biomarkers with genetic algorithm and principal component analysis[J]. BMC Bioinformatics, 2019, 20(Suppl 6): 413.
[20] National Medical Products Administration. Notice on the application of bezoar and its substitutes[N/OL].(2004-01-21) [2021-04-21]. https://www.nmpa.gov.cn/xxgk/fgwj/gzwj/gzwjyp/20040121010101169.html.

基于机器学习鉴别牛黄类药材红外光谱的研究

Succession medicinal substances of calculus bovis with infrared spectroscopy coupled with machine learning methods

PDF (PC)

可视化

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 15

编辑推荐

Metrics

[1]	贾茵茵, 张财树, 周颖, 刘丽, 耿颖, 魏宁漪, 钮思静, 陈华, 许卉. 吸入制剂微细粒子空气动力学粒径分布评价方法研究[J]. 中国药物警戒, 2024, 21(2): 141-146.
[2]	高云娟, 赵旭, 白天凯, 柏兆方, 王伽伯, 宋海波, 肖小河. 基于不良反应监测大数据的药品安全风险发现与识别策略[J]. 中国药物警戒, 2024, 21(1): 1-5.
[3]	陈枭, 郭晓晶, 许金芳, 韦连慧, 陈晨鑫, 梁际洲, 郑轶, 叶小飞. 机器学习方法在FAERS布加替尼不良反应信号检测中的应用[J]. 中国药物警戒, 2023, 20(6): 639-645.
[4]	郑轶, 罗枭, 张朋朋, 刘永梅, 叶小飞, 郭晓晶, 陈霞, 贺佳. 基于主动监测系统的免疫检查点抑制剂心肌炎不良反应探索[J]. 中国药物警戒, 2023, 20(6): 634-638.
[5]	黄茵茵, 张立雯, 刘静, 陈静, 周远华, 许明哲, 姚尚辰, 王晨, 张现化. 普卢利沙星3种晶型的制备、表征与稳定性研究[J]. 中国药物警戒, 2023, 20(4): 374-378.
[6]	耿颖, 刘毅, 王雪蕾, 谭德讲, 朱容蝶, 郭贤辉, 陈华, 郑丽娥. 近红外光谱法和统计学过程控制方法对亚叶酸钙药品对照品水分进行质量控制的研究[J]. 中国药物警戒, 2023, 20(11): 1221-1227.
[7]	彭静, 庹亚莉, 李巧玲, 周荣, 李新林, 彭惠. 基于国外药物警戒安全性自发呈报系统的氯巴占儿童不良事件数据分析[J]. 中国药物警戒, 2023, 20(11): 1264-1268.
[8]	王涛, 郑明节, 刘红亮, 王青, 沈传勇. 人工智能在美国药物警戒中的应用现状及启示[J]. 中国药物警戒, 2023, 20(10): 1129-1133.
[9]	黄露, 刘博, 王岩, 范慧红, 张庆生. 核苷类抗病毒药物质量控制研究进展[J]. 中国药物警戒, 2022, 19(6): 692-696.
[10]	欧阳罗丹, 肖苏萍, 周建新, 陈蒙, 邓哲, 周海燕, 杜杰, 王继永. 桂枝的薄层鉴别及含量测定[J]. 中国药物警戒, 2022, 19(4): 380-383.
[11]	朱兰, 朱彦, 姚克宇, 陈超, 刘丽红. 中药药源性肝损伤文献个案报道数据元素分析[J]. 中国药物警戒, 2022, 19(12): 1352-1357.
[12]	石岩, 魏锋, 马双成. 基于RGB多通道图像数据的牛胆与羊胆薄层色谱特征分析[J]. 中国药物警戒, 2022, 19(10): 1060-1066.
[13]	赵荣华, 郭姗姗, 孙静, 时宇静, 包蕾, 耿子涵, 鲍岩岩, 周冠儒, 高英杰, 崔晓兰, 王道涵. 体外培育牛黄对人冠状病毒肺炎湿毒疫模型的机制研究[J]. 中国药物警戒, 2021, 18(12): 1117-1122.
[14]	周明霞, 高会芹, 彭雄颖, 李亚涛, 李军山, 屈云萍, 李振江. 地黄不同炮制品配方颗粒的红外快速鉴别[J]. 中国药物警戒, 2017, 14(2): 87-89.
[15]	顾冰. 89例牛黄解毒片/丸不良反应文献分析[J]. 中国药物警戒, 2016, 13(6): 359-363.