基于科学计量的机器学习技术发展评估分析

机器学习是通过数据训练使机器获取新的知识和技能的计算机技术，是人工智能技术的核心和前沿领域之一。本文以Web of Science?核心合集收录的机器学习领域相关SCI论文和incoPat全球专利数据库收录的机器学习领域相关专利为数据源，运用科学计量方法对文献数据和专利数据进行时间、地域和机构分布分析，展示了机器学习领域研究实力的分布情况，并对机器学习的发展趋势进行了分析；然后利用主题聚类分析方法及可视化软件VOSviewer，挖掘出机器学习领域的关键技术、技术热点；最后对结果进行了总结分析，以期为我国机器学习发展布局提供参考和借鉴。

Development analysis of machine learning technology based on scientometrics

Machine learning, as a core and pioneering technology in artificial intelligence, is a computer technology which enables machines to acquire new knowledge and skills with data training. This paper takes the SCI papers from the Web of ScienceTM and the patents from the incoPat global patent database as data source, and utilizes scientometrics as the analysis method and visualization software as the analysis tool. The quantity, subject area distribution, research institutions are analyzed, showing the distribution of research strength in machine learning. Using the topic cluster analysis method and information visualization software VOSviewer, the key technologies and technical hotspots of machine learning are excavated, and the development trend of machine learning is predicted. Finally, the results are summarized and analyzed, providing references for the overall arrangement of machine learning development in China.