Double clustering analysis of medical artificial intelligence research hotspots

Objective To analyze the international research results of artificial intelligence in the medical field by the double clustering method, and to explore the hot trends in the topic field.

Methods The Web of Science core collection database was searched for the research literature of artificial intelligence in the field of medicine, and the high-frequency keywords were extracted by Co-Occurrence13.4 to generate the word matrix. The gCluto1.0 clustering toolkit was used for the double cluster analysis.

Results A total of 7 803 articles were included, and the annual number of publications showed an overall upward trend. The United States ranked the first in the total number of publications. A total of 30 high-frequency subject words were extracted to form 6 clusters such as artificial intelligence applied to biomarker detection. The research hotspots focused on six topics: health care, disease outcome, whole-course disease monitoring, auxiliary diagnosis of cancer, model validity and differential biomarkers.

Conclusion Artificial intelligence has been widely used in clinical diagnosis and treatment technology, which provides targeted support for genetic testing and public health events. However, related domestic research is still in developing stage. In the future, we need to rely on multidisciplinary and inter-institutional communication and cooperation to promote the development of intelligent medical in China, so that it truly becomes an important tool to promote the development of medical and health services.