基于近红外光谱结合化学计量学方法的山里红产地溯源分析

目的 基于近红外光谱（near-infrared spectroscopy，NIRS）结合机器学习算法模型，建立山里红Crataegus pinnatifida var. major的产地溯源技术。方法 收集6个省份的91份山里红样本，采集其NIRS，应用多种机器学习算法，包括主成分分析（principal component analysis，PCA）、正交偏最小二乘法判别分析（orthogonal pmjartial least squares-discriminant analysis，OPLS-DA）、K-最近邻（K-nearest neighbor，KNN）、决策回归树（classification and regression tree，CART）、随机森林（random forest，RF）、朴素贝叶斯（naive bayes，NB）、线性判别分析（linear discriminant analysis，LDA）和神经网络（artificial neural network，ANN）算法，探讨适合山里红产地溯源的模型。结果 ANN模型的准确率和模型稳定性最优，可作为山里红产地识别的首选模型。结论 NIRS结合ANN模型是山里红产地溯源的有效手段，为山里红的产地溯源提供了科学参考。

Geographical origin traceability of Crataegus pinnatifida var. major based on near-infrared spectroscopy combined with chemometrics

Objective To develop a traceability technology system for determining the origin of Crataegus pinnatifida var.major through the integration of near-infrared spectroscopy (NIRS) technology and machine learning algorithms. Methods A total of 91 samples of C. pinnatifida var. major were collected from six provinces in China, and their NIRS were acquired. Various machine learning algorithms, including principal component analysis (PCA), orthogonal partial least squares-discriminant analysis (OPLS-DA), k-nearest neighbor (KNN), classification and regression tree (CART), random forest (RF), naive bayes (NB), linear discriminant analysis (LDA) and artificial neural network (ANN), were employed to establish a model for the purpose of origin tracing. Results Among the different algorithms tested, the ANN model demonstrated the highest accuracy and stability in identifying the origin of C. pinnatifida var. major, making it a reliable tool for traceability. Conclusion The combination of NIRS technology and the ANN model can be used as an effective approach for tracing the geographical origin of C. pinnatifida var. major. This study contributes to the establishment of a scientifically rigorous foundation for the geographical origin tracing of C. pinnatifida var. major.