A COMPREHENSIVE EVALUATION OF MACHINE LEARNING ALGORITHMS FOR PREDICTING DIABETES ONSET AND ITS CLINICAL APPLICATIONS

Abstract [English]

The early prediction of diabetes onset is critical in preventing the progression of the disease and improving patient outcomes. This research presents a comprehensive evaluation of various machine learning algorithms applied to the prediction of diabetes onset, with a focus on their clinical applications. A range of machine learning models, including decision trees, random forests, support vector machines, and neural networks, are assessed using a publicly available diabetes dataset. The study aims to identify the most accurate and efficient algorithms for predicting diabetes onset based on a set of medical attributes, such as age, BMI, glucose levels, and family history. The evaluation considers multiple performance metrics, including accuracy, precision, recall, and area under the curve (AUC), to assess the effectiveness of each algorithm in predicting diabetes risk. The results demonstrate that ensemble methods like random forests and gradient boosting outperformed other models, providing high accuracy and robustness in prediction. Additionally, the study discusses the practical implications of these models in clinical settings, highlighting their potential for aiding healthcare providers in early diabetes detection and personalized treatment plans. The findings contribute valuable insights into the use of machine learning for diabetes prediction, offering a foundation for future research and the development of automated decision support systems in healthcare.