郑州大学学报（医学版）2025，Vol.60Issue(5)：618-622,5.DOI:10.13705/j.issn.1671-6825.2025.04.196
山茱萸产地LIBS结合多模型融合判据鉴别方法的建立
Construction of cornus officinalis origin identification method using la-ser-induced breakdown spectroscopy combined with multi-model fusion criterion
摘要
Abstract
Aim:To establish a method for identifying the origin of cornus officinalis using laser-induced breakdown spectroscopy(LIBS)combined with multi-model fusion criterion.Methods:Dried mature pulp of cornus officinalis genus from 5 different origins was analyzed using LIBS.Principal component analysis(PC A)was employed for feature dimension-ality reduction.Four machine learning algorithms including support vector machine(SVM),random forest(RF),extreme gradient boosting(XGBoost),and backpropagation neural network(BPNN)were optimized via Bayesian hyperparameter tuning to establish identification models.A weighted ensemble method was applied to integrate the results of these 4 models through normalized weighting,constructing a multi-model fusion criterion approach.Results:Eleven principal components was identified by PC A as the feature vectors,4 machine learning classification models achieved accuracies of 91.00％,93.00％,93.50％,and 94.00％,respectively,in discriminating the geographical origins.The multi-model fusion method a-chieved an accuracy of 96.00％,with high sensitivity(0.900 0-1.000 0),precision(0.930 0-1.000 0),and F1 score(0.940 0-0.980 0)across samples from different origins.Conclusion:The combination of LIBS and multi-model fusion could efficiently and accurately identify the cornus officinalis's origins.关键词
激光诱导击穿光谱/机器学习/加权融合/山茱萸/产地鉴别Key words
laser-induced breakdown spectroscopy/machine learning/weighted ensemble/cornus officinalis/origin i-dentification分类
化学化工引用本文复制引用
吕斌,黄现青,吕文浩..山茱萸产地LIBS结合多模型融合判据鉴别方法的建立[J].郑州大学学报（医学版）,2025,60(5):618-622,5.基金项目
河南省高校科技创新团队支持计划项目(23IRTSTHN023) （23IRTSTHN023）
