[关键词]
[摘要]
目的 基于文献计量学与可视化知识图谱,系统分析人工智能(artificial intelligence,AI)赋能中药智能制造的研究现状、热点和发展趋势,为后续研究提供参考。方法 检索中国知网、万方、维普、SinoMed、Web of Science和PubMed等数据库中相关文献。经去重筛选后,用VOSviewer和CiteSpace软件对发文趋势、作者、机构、期刊、关键词等进行可视化分析。结果 共纳入中文文献803篇、英文文献247篇,整体发文量逐年上升。中英文文献发文量最多的作者分别为乔延江和Ren Xiaoliang。中英文文献发文量最多的机构分别为北京中医药大学和天津中医药大学,各研究团队内部合作紧密,但团队间合作较少。发文量最多的中英文期刊分别为《中国中药杂志》和Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy。当前研究热点主要为机器学习、深度学习、卷积神经网络等AI技术在中药识别、质量评价、工艺优化与质量控制等场景中的应用。技术路径经历了从“浅层模型+单模态”到“深度学习+多模态融合”的演进过程。结论 近年来AI赋能中药智能制造领域发展迅速,AI已逐渐成为驱动中药智能制造转型升级的重要力量。未来需加强多模态数据建设与共享,推进可解释AI建模与全过程智能化,强化跨学科和跨团队合作,健全标准化监管体系,以推动中药工业高质量发展。
[Key word]
[Abstract]
Objective To systematically analyze the current research landscape, hotspots, and development trends in the field of artificial intelligence (AI)-empowered intelligent manufacturing of traditional Chinese medicine (TCM) based on bibliometrics and visualized knowledge graphs, providing a reference for subsequent research in this field. Methods Relevant articles were retrieved from the China National Knowledge Infrastructure (CNKI), Wanfang, VIP, SinoMed, Web of Science, and PubMed databases. After deduplication and screening, VOSviewer and CiteSpace software were employed for visual analysis of publication trends, authors, institutions, journals, keywords, etc. Results A total of 803 Chinese literature and 247 English literature were included. Overall publication volume showed a year-on-year upward trend. The authors with the highest numbers of publications in the Chinese and English literature were Qiao Yanjiang and Ren Xiaoliang, respectively, while the institutions with the highest numbers of publications were Beijing University of Chinese Medicine and Tianjin University of Traditional Chinese Medicine, respectively. Intra-team collaboration was close among individual research groups, whereas inter-team connections remained relatively limited. The Chinese and English journals with the highest numbers of publications were the China Journal of Chinese Materia Medica and Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy, respectively. The current research hotspots mainly lie in the application of AI technologies such as machine learning, deep learning, and convolutional neural network in scenarios such as the identification, quality evaluation, process optimization, and quality control of TCM. The technical path has undergone an evolution process from “shallow model + single modality” to “deep learning + multi-modal fusion”. Conclusion In recent years, the field of AI-empowered intelligent manufacturing of TCM has developed rapidly, with AI progressively emerging as a core driving force behind the transformation and upgrading of TCM intelligent manufacturing. Future efforts should focus on strengthening multi-modality data construction and sharing, promoting explainable AI modeling and entire process intelligence, enhancing interdisciplinary and cross-team collaboration, and improving standardized supervision system, with the aim of advancing high-quality development of the TCM industry.
[中图分类号]
R282
[基金项目]
国家科技重大专项(2026ZD1806400); 上海市领军人才计划项目(SHLJ2019100); 上海中医药大学科技发展项目(标准化专项)(24BZH16)