Abstract : Diabetes mellitus is a prevalent metabolic disorder globally. Its primary etiologies encompass socioeconomic determinants, behavioral risk factors, and underlying comorbidities. Numerous epidemiological studies have investigated various diabetes phenotypes, impacting both sexes across the entire age spectrum. This study utilizes a dataset containing clinical profiles of 1,000 subjects assessed on multiple biometric and sociodemographic variables. The objective is to classify diabetes into type 1, type 2, and prediabetes using an array of deep learning and machine learning algorithms. Currently, artificial intelligence-driven diagnostic methods represent a state-of-the-art approach for disease stratification. This research evaluates the performance of six classification algorithms for determining glycemic status: random forest (RF), support vector machine (SVM), extreme gradient boosting (XGBoost), convolutional neural network (CNN), recurrent neural network (RNN), and long short-term memory (LSTM) network. Results demonstrate that the XGBoost classifier attained the highest predictive accuracy of 91% with a training duration of 20 seconds, surpassing the other models. These findings underscore the potential of advanced computational algorithms for precise diabetes phenotyping and risk assessment, offering significant implications for disease management and public health interventions.
M. et al. (Fri,) studied this question.