A Review on AI-Driven Approache for Diabetes Prediction in Female Populations via Risk Factor Analysis

Abstract

Diabetes is fast emerging as a serious health  challenge in the world, with countries around the globe spending huge resources on its management. At the same time it impacts  the quality of life, especially among the female population in  view of the hormonal, reproductive, metabolic and lifestyle related risk factors. Early diagnosis and timely treatment are  essential to minimize complications and enhance preventive  healthcare. The present review paper presents the different  artificial intelligence (AI) based strategies for the prediction of  diabetes based on risk factors. The study focuses on the female specific risk factors like gestational diabetes mellitus (GDM),  polycystic ovary syndrome (PCOS), hormonal imbalance, obesity, genetic predispositions and socioeconomic factors  related to the progression of diabetes. Furthermore, the survey  examines how predictive accuracy can be improved by using  machine learning (ML), deep learning (DL), federated learning,  and cloud-based healthcare systems and how they can be used  to personalize healthcare and facilitate early healthcare  diagnosis. The results show that AI-based models, such as  multimodal and DL models, outperform the traditional models in terms of prediction accuracy. However, there are a number  of challenges that are still important research challenges for AI  predictive systems for diabetes, such as limited data  interpretability, privacy concerns, imbalanced data, and lack of female-specific data.