基于CiteSpace的人工智能领域机器学习算法研究的可视化探析
Visual Analysis of the Research on Machine Learning Algorithms in the Field of Artificial Intelligence Based on CiteSpace

人工智能定义了智能系统的终极愿景，机器学习是人工智能的技术基石。本文以“人工智能”和“机器学习”为主题，对中国知网核心数据库进行检索，遴选出1992~2024年2382篇文献作为研究样本，借助CiteSpace 6.4.R1可视化软件，采用共现分析、聚类分析以及突现词分析等文献计量方法，综合梳理人工智能领域机器学习算法研究的热点、发展脉络和演进历程。分析结果表明，文献发表数量在2016年后大幅增加。研究机构以中国科学院大学和清华大学为核心形成两大科研机构合作网络，但总体来说，研究机构之间合作仍显有限，呈现出“小团体”合作的特征。研究热点主要为人工智能、机器学习、深度学习、大数据、神经网络等，深度学习革命带来训练效率提升，大数据推动算法迭代，“数据–算法–算力”是该领域发展的核心驱动力。此外，通过共现时区图和突现词分析发现，未来研究的主要发展方向聚焦于大模型、预后预测、可解释性、数据共享等领域，学者开始关注该领域的隐私保护与伦理问题。应用场景上，在算法行政、智慧医疗、自动驾驶、智能决策等方面进一步深入研究，研究重点从以理论和算法优化为核心的基础研究逐步转向技术优化、应用落地及合理使用，进入人工智能领域机器学习算法研究新时代。
Artificial intelligence defines the ultimate vision of intelligent systems, and machine learning serves as the technological cornerstone of artificial intelligence. Taking “artificial intelligence” and “machine learning” as the themes, this paper conducts a search in the core database of the China National Knowledge Infrastructure (CNKI). A total of 2,382 papers published from 1992 to 2024 are selected as the research samples. With the help of the CiteSpace 6.4.R1 visualization software and using bibliometric methods such as co-occurrence analysis, cluster analysis, and burst term analysis, this paper comprehensively sorts out the research hotspots, development context, and evolution process of machine learning algorithms in the field of artificial intelligence. The analysis results show that the number of published papers increased significantly after 2016. The research institutions, with the University of Chinese Academy of Sciences and Tsinghua University at the core, have formed two major cooperation networks among scientific research institutions. However, overall, the cooperation among research institutions is still limited, showing the characteristics of “small-group” cooperation. The research hotspots mainly include artificial intelligence, machine learning, deep learning, big data, and neural networks. The deep-learning revolution has improved training efficiency, and big data has promoted algorithm iteration. “Data, algorithm, and computing power” are the core driving forces for the development of the field. In addition, through the analysis of the co-occurrence time-zone map and burst terms, it is found that the main future development directions of research focus on large models, prognosis prediction, interpretability, and data sharing. Scholars have begun to pay attention to privacy