Introduction to DRPM

The DRPM is a deep learning model that utilizes data from the National Health and Nutrition Examination Survey (NHANES) to predict the risk of diabetes. To obtain more effective features and provide sufficient information for diabetes risk identification, this study employed a feature fusion and extraction mechanism. Subsequently, an optimal feature subset was composed of five features: FBG, age, WtHR, MSP, and BMI. The final model was constructed based on these five features. The model achieved Area Under the Curve (AUC) scores of 0.9399, 0.9358, and 0.9601 for the training, validation, and testing datasets, respectively.The AUC score on the independent validation set was 0.93. These findings indicate that the predictive model developed in this study demonstrates a certain level of identification ability for the diabetic population.