Abstract
Context: The diagnostic methods for diabetes mellitus (DM), a chronic metabolic disorder characterized by elevated blood sugar levels, are rapidly evolving thanks to artificial intelligence (AI), particularly machine learning (ML) and deep learning (DL). This review explores the applications of AI in risk assessment and diagnosing different types of diabetes. Evidence Acquisition: The review highlights the effectiveness of various ML models, including support vector machines (SVMs), random forests (RFs), and DL techniques like convolutional neural networks (CNNs), in achieving high diagnostic accuracy. Challenges include limited data availability, interpretability of complex models, and the need for standardized performance metrics. Results: Machine learning methods like SVMs and RFs are highly effective at diagnosing different types of diabetes, and DL techniques like CNNs also show great promise. Conclusions: Overall, AI has immense potential to revolutionize diabetes diagnosis by facilitating risk assessment and early detection, improving treatment efficacy, and preventing severe complications.