Optimizing Prediabetes Diagnosis Through Knowledge-Based Systems

Abstract

The escalating global prevalence of prediabetes highlights the urgency of preventive measures, particularly given its association with increased age, obesity, and additional risk factors. Addressing this concern, the explainability component of Artificial Intelligence (AI) emerges as a valuable asset in diabetes prevention strategies. This study adopts an experimental design grounded in knowledge-based systems, utilizing the knowledge engineering method to craft a web-based health tool for diabetes diagnosis. The process encompasses acquisition, representation, validation, inferencing, and explanation phases. The online diagnostic tool not only facilitates self-diagnosis but also delivers conclusive findings and enables user registration. Practical solutions and preventive recommendations are offered, aligning with the overarching goal of diabetes prevention. The study identifies three operational phases – self-diagnosis, presentation of final findings, and member registration. To enhance the application's efficacy, the analysis provides constructive suggestions for future refinements and advancements. This research underscores the potential of AI-driven, explainable systems in contributing to the global effort to combat the rising prevalence of diabetes.