目的 利用多源智能感官信息融合技术建立中药寒热预测模型，为中药寒热药性评价提供新的方法学参考。方法 根据《中国药典》2020年版一部中性味与归经筛选出97种具有寒热属性（寒、微寒、凉为寒；温、微温、热为热）的中药，以所选样品寒热药性作为标签信息（Y）；通过PEN3型电子鼻、SA402B型电子舌、ASTREE型电子舌采集上述样品的传感器数据矩阵为自变量（X）；分别利用判别分析（discriminant analysis，DA）、主成分分析-判别分析（principal component analysis-discriminant analysis，PCA-DA）、偏最小二乘-判别分析（partial least squares-discriminant analysis，PLS-DA）、支持向量机（support vector machine，SVM）、最小二乘支持向量机（least square support vector machine，LS-SVM）5种算法建立Y＝F(X)关系的单源、多源分类辨识模型，以正判率为指标优选模型。结果 经留一法交互验证，寒热二分类基于单源信息融合最优模型正判率为81.44%（PEN3/PLS-DA），基于多源信息融合辨识最优模型正判率为85.57%（ASTREE＋SA402B/LS-SVM），多源信息融合后模型的正判率有所提高。结论 该研究利用多源智能感官信息融合技术可以在一定程度上提高中药寒热药性分类的辨识，为进一步将其应用于中药药性现代化的研究提供新的参考。

Objective To establish a cold and heat prediction model of traditional Chinese medicine by using multi-source intelligent sensory information fusion technology, and provide a new methodological reference for the evaluation of cold and heat properties of traditional Chinese medicine. Methods According to the neutral flavor and meridian tropism of the 2020 edition of the Pharmacopoeia of the People’s Republic of China, 97 kinds of traditional Chinese medicines with cold and hot properties (cold, slightly cold, and cool is classified as cold; warm, slightly warm, and hot is classified as hot) were selected, the cold and hot medicinal properties of the selected samples were used as label information (Y). The sensor data matrix of the above samples collected by PEN3 electronic nose, SA402 B electronic tongue and ASTREE electronic tongue was the independent variable (X). Discriminant analysis (DA), principal component analysis-discriminant analysis (PCA-DA), partial least squares-discriminant analysis (PLS-DA), support vector machine (SVM) and least squares support vector machine (LS-SVM) were used to establish the single-source and multi-source classification identification models of Y = F(X) relationship, and the model was optimized with the positive discrimination rate as the index. Results Through the leave-one-out cross-validation method, the correct rate of the optimal model based on single-source information fusion is 81.44% (PEN3/PLS-DA), and the correct rate of the optimal model based on multi-source information fusion is 85.57% (ASTREE + SA402B/LS-SVM). The correct rate of the model after multi-source information fusion is improved. Conclusion In this study, multi-source intelligent sensory information fusion technology can improve the identification of cold and heat property classification of traditional Chinese medicine to a certain extent, and provide a new reference for further application in the modernization of traditional Chinese medicine.

R283.6

河南省自然科学基金杰出青年科学基金项目（242300421023）；河南省卫生健康中青年学科带头人专项（HNSWJW-2020014）；2022年青年岐黄学者培养项目（[2022]256）；河南省卫生健康委员会国家中医临床研究基地科研专项课题（2021JDZY106）；河南省卫生健康委国家中医临床研究基地科研专项（2021JDZY014）