Artificial Intelligence in Diabetes Management in Last Two Decades: A Bibliometric Analysis on Web of Science

Abstract

Objectives: To conduct a bibliometric analysis to analyze the landscape of artificial intelligence(AI) in diabetes management in the past two decades.Methods: The literatures on AI in diabetes management during 2003 to 2021 were respectively retrieved from Web of Science database. CiteSpace software was applied to perform analysis of frequency distribution and co-occurrence of sub-keyword terms, frequency distribution analysis of reference citations, and analysis of burst detection methods for keyword terms and cited references, in order to create a knowledge map to reveal diabetes management hotspots and cutting-edge directions.(p/)(p)Results: A total of 1,242 articles on diabetes AI managements were retrieved. The results showed that the number of papers increased year by year, and the number of papers published was the largest in 2019-2021, suggesting that in the field of diabetes management, digital technology was widely adopted in the daily management of diabetes and was conducive in early screening for diabetes complications. In the field of AI in diabetes, various machine learning models including latest deep learning have been proposed and investigated. In the literatures on diabetes digital interventions, the keywords with higher frequency were management, intervention, diabetes, care and blood sugar control. Comparatively, the keywords with the highest frequency in diabetes AI are machine learning, management, diabetes, systems and classification.Conclusions: Publication outputs of the application of AI in DM management have been rising over last two decades. Some hotspots such as mobile medicine, telemedicine, deep learning, and digital therepeutics are of great significance for future research.

Ähnliche Arbeiten