关键词: 文献计量学, 机器学习, 医学, 学位论文, 研究热点

Abstract: Objective To find and analyze the research trends and future research directions of academic dissertations of medical machine learning in China. Methods All academic dissertations on medical machine learning published in full text in Chinese were obtained by searching the CNKI and Wanfang database. The retrieval period was from the establishment of the database to February 14, 2023. The search keywords included ″medicine″ and ″machine learning″. Microsoft Excel 2016 software was used to organize data from retrieved records, including the year of thesis inclusion, institution of degree, supervisor and cited frequency. VOS viewer software was used to visualize and analyze the high-frequency keywords. Results A total of 2 662 academic dissertations in the field of machine learning were retrieved from 299 universities. The combined number of dissertations from the top 10 universities accounted for 29.9%(797/2 662) of all the dissertations. There were 2 533 academic dissertation supervisors, and the citation frequency of the dissertations ranged from 0 to 204 times. However, 46.2%(1 229/2 662) of the published dissertations had never been cited. The research topics of the dissertations mainly included assisted diagnosis in medical imaging, extraction of electronic medical record texts, optimization of medical image segmentation algorithms, and efficient and accurate utilization of multimodal models. Conclusions There is a relatively large number of machine learning-related dissertations in the field of medicine, as well as a significant number of institutions and supervisors. However, the quality of the dissertations needs improvement, and the research hotspots are relatively concentrated in the area of medical assisted diagnosis.

Key words: Bibliometrics, Machine learning, Medicine, Dissertation, Research hotspots