AI-Powered Innovations in Diabetes: A Narrative Review

Abstract

Artificial intelligence (AI) has been increasingly used in diabetes care with promising results of improved outcomes. This is a narrative review on the application of AI in diabetes mellitus (DM) management in Malaysia. Bibliographic search with the search terms of “diabetes mellitus”, “artificial intelligence”, “machine learning” and “Malaysia” was conducted on PubMed, Scopus, and Google Scholar. A final list of 65 publications were included for analysis in this review. Most of the studies (n=28, 43.1%) were done with the focus on DM in general. The types of AI most employed by the studies were neural network (n=15, 23.1%), supervised learning together with neural network (n=13, 20.0%), and supervised learning (n=12, 18.5%). AI was most applied in the classification (n=15, 23.1%), prediction (n=15, 23.1%), detection (n=11, 16.9%), diagnosis (n=9, 13.8%), and identification (n=7, 11.0%) of DM. High levels of accuracy, sensitivity, specificity, and precision (more than 90%) were reported in the included studies.

Ähnliche Arbeiten