Overcoming Data Scarcity: Strategic Data Manipulation in the Development of a Knowledge-Based System for Type 2 Diabetes Prediction

The effective prediction of Type 2 Diabetes (T2D) risk requires the integration of both clinical data and social determinants of health. However, the scarcity of comprehensive datasets that include relevant social determinants presents significant challenges. This paper focuses on the critical data manipulation processes undertaken to address these challenges in the development of a flexible and adaptive Knowledge-Based System (KBS) for T2D prediction. Through systematic data refinement and the strategic exclusion of irrelevant attributes, the KBS was designed to accommodate new data as it becomes available, ensuring its ongoing relevance and effectiveness. The potential versatility of the KBS is further supported by cited case studies from existing literature, demonstrating its applicability across diverse socio-economic and geographic contexts. This research highlights the importance of robust data manipulation techniques in overcoming data scarcity and future-proofs the KBS to adapt to evolving public health needs.