Quantitative Analysis of Disease Dynamics in Machine Learning Models for Diabetes Prediction

Abstract

Diabetes is a prevalent chronic disease that poses significant health risks worldwide. Early prediction of diabetes can improve patient outcomes by enabling timely interventions. This study evaluates the performance of various machine learning (ML) models, including Random Forest, Support Vector Machine (SVM), Logistic Regression, and Decision Trees, to predict diabetes based on clinical features. The analysis focuses on metrics such as accuracy, precision, recall, F1 score, and AUC-ROC, with Random Forest and SVM showing superior predictive power. However, Logistic Regression and Decision Trees offer greater interpretability, a crucial factor for clinical applications where model transparency is essential. Feature importance analysis identified glucose level and BMI as key predictors, aligning with clinical insights and supporting the models' relevance. The findings suggest that ML models, particularly Random Forest, could be instrumental in diabetes screening, though further research with diverse datasets and explainable AI techniques is recommended to enhance model generalizability and clinical trust. This study contributes to the field of predictive healthcare by underscoring the potential of ML models to support diabetes prevention and management through early identification of high-risk individuals.