Comparative Analysis of Machine Learning Models for Diabetes Prediction

Abstract

Diabetes is a chronic health condition affecting millions worldwide, and early detection plays a vital role in effective disease management and prevention. In this study, we conduct a comparative analysis of four machine learning models—Logistic Regression, Random Forest, Gradient Boosting, and Linear Regression—applied to the Pima Indian Diabetes dataset obtained from Kaggle. The dataset comprises diagnostic measurements of female patients aged 21 and above of Pima Indian heritage. Each model is evaluated using key classification metrics, including accuracy, precision, recall, and F1-score. Among the models, Logistic Regression and Gradient Boosting achieved the highest accuracy of 75%, while Random Forest and Linear Regression showed slightly lower performance at 72% and 73.16%, respectively. The study highlights the effectiveness of ensemble methods and traditional classifiers in predicting diabetes outcomes and provides insight into their relative strengths for clinical decision support systems. These results suggest that machine learning can be a valuable tool in aiding early diagnosis and improving patient care strategies.