Using Logistic Regression and Bayesian Regression in The analysis of Chronic Disease Data

Abstract

This study aims to evaluate the effectiveness of Bayesian modeling in analyzing large-scale medical data by employing the Bayesian logistic regression model and comparing its results with those of the classical logistic regression model. The research relied on a dataset of 50,000 public health records containing demographic and medical information of patients, including those suffering from chronic diseases such as diabetes and heart disease. The dependent variable (Y) represented the presence or absence of a chronic disease, while the independent variables included age, body mass index (BMI), blood pressure, and blood sugar.

The results revealed that the Bayesian model achieved a classification accuracy of 86.7%, compared to 63.2% for the classical model. Furthermore, the Bayesian model recorded a higher AUC value (0.91) versus 0.87 for the classical approach, highlighting the superiority of Bayesian inference in handling large datasets and expressing uncertainty through probability distributions. These findings emphasize that Bayesian modeling offers a more flexible and accurate statistical framework compared to traditional methods, making it an effective tool in supporting future medical decision-making.