A Comparative Study of Machine Learning Models for Predicting Chronic Diseases

Abstract

In this study, we explore the predictive capabilities of various machine learning models in identifying two major chronic conditions: diabetes and heart disease. Using publicly available datasets, we trained and evaluated Logistic Regression (LR), Random Forest (RF), K-Nearest Neighbors (KNN), and Support Vector Machine (SVM) models. Performance was assessed using Accuracy and ROC-AUC metrics. For heart disease, the best-performing model was Random Forest with an accuracy of 0.88. For diabetes, Logistic Regression performed best with an accuracy of 0.75. Our findings reinforce the value of ML in preventive healthcare and suggest promising directions for future improvements.