부산대학교 도서관

This research pioneers the integration of Keras TensorFlow to engineer a cutting-edge machine-learning model tailored for the precise prediction of Diabetic Foot Ulcers (DFUs) within a binary classification framework, differentiating between Diabetes Mellitus (DM) and Control groups. Employing a meticulous confusion matrix analysis, the model's performance metrics unveil remarkable outcomes: True Positives (TP) = 91, False Positives (FP) = 0, False Negatives (FN) = 9, and True Negatives (TN) = 92. The model exhibits a stellar accuracy of 98.26. The strategic adoption of Keras TensorFlow stems from its established reputation as a versatile and powerful deep-learning library. Noteworthy is the model's adeptness in accurately identifying DFUs within the DM group and effectively discerning non-DFU cases in the Control group, underscoring the intrinsic efficacy of Keras TensorFlow for precision-centric tasks. This study represents a substantial leap forward in the precision medicine paradigm within diabetes care, showcasing the transformative potential of a meticulously crafted machine-learning model for enhancing DFU prediction. The results obtained underscore the applicability of Keras TensorFlow in elevating diabetic care strategies through early and precise detection of foot ulceration, positioning this innovative approach at the forefront of precision medicine in diabetes management.