数据融合技术是对多来源的数据信息进行统筹、集成的技术，以提高决策模型的灵敏度、特异性和准确率。数据融合技术结合多变量模型是研究中药复杂体系的有力工具，目前已应用于中药的基原鉴别、产地溯源与鉴别、质量控制与评价、加工炮制与制剂研究、资源形成研究等诸多领域。融合所纳入的源数据主要是中药的化学物质信息，包括各类色谱、光谱信息，无机元素及有机成分的含量信息，电子鼻、电子眼、电子舌等传感器信息以及代谢组学信息等；所采用的多变量模型如主成分分析、层次聚类分析、偏最小二乘-判别分析、正交偏最小二乘-判别分析、支持向量机、人工神经网络、随机森林、决策树、线性判别分析等。未来，数据融合有望与人工智能（artificial intelligence，AI）相结合，源数据中纳入生物医学数据与组学数据，拓展应用于中药活性物质筛选，预测患者对药物反应、药物相互作用、药物-靶点相互作用，中药新药开发以及栽培种植等更多领域，同时应积极开发集数据融合与多变量建模功能于一体的中药研究相关软件系统。

Data fusion is a technology that coordinates and integrates data information from multiple sources to improve the sensitivity, specificity and accuracy of decision models. Data fusion combined with multivariable model is a powerful tool to study the complex system of traditional Chinese medicine (TCM), which has been applied to many research fields such as species identification, origin tracing and identification, quality control and evaluation, processing and preparation, and resources formation research. The source data included in the fusion is mainly chemical substance information of TCM, including various types of chromatographic and spectral information, content information of inorganic elements and organic components, sensor information such as e-nose, e-eye and e-tongue, and metabolomics information, etc. The multivariate models used are principal component analysis (PCA), hierarchical cluster analysis (HCA), partial least squares-discriminant analysis (PLS-DA), orthogonal partial least squares-discriminant analysis (OPLS-DA), support vector machine (SVM), artificial neural network (ANN), random forests (RF), decision trees, linear discriminant analysis (LDA), etc. In the future, data fusion is expected to be combined with artificial intelligence (AI), incorporate biomedical and omics data into source data, and expand applications in more fields such as screening of active substances in TCM, prediction of patient’s response to drug, drug-drug interactions, drug-target interactions, development of new TCM, and cultivation of TCM, etc. At the same time, software systems related to TCM research that integrate data fusion and multivariate modeling function should be actively developed.

R28；TP391

甘肃省青年科技基金计划项目（21JR7RA634）；甘肃省自然科学基金资助项目（20JR5RA154）